transformers 5.0.0__py3-none-any.whl → 5.0.0rc0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (1606) hide show
  1. transformers/__init__.py +36 -55
  2. transformers/activations.py +1 -1
  3. transformers/audio_utils.py +33 -32
  4. transformers/cache_utils.py +139 -32
  5. transformers/cli/chat.py +3 -3
  6. transformers/cli/serve.py +19 -49
  7. transformers/cli/transformers.py +1 -2
  8. transformers/configuration_utils.py +155 -129
  9. transformers/conversion_mapping.py +22 -158
  10. transformers/convert_slow_tokenizer.py +17 -227
  11. transformers/core_model_loading.py +185 -528
  12. transformers/data/data_collator.py +4 -12
  13. transformers/data/processors/glue.py +1 -0
  14. transformers/data/processors/utils.py +1 -0
  15. transformers/data/processors/xnli.py +1 -0
  16. transformers/dependency_versions_check.py +1 -0
  17. transformers/dependency_versions_table.py +7 -5
  18. transformers/distributed/configuration_utils.py +2 -1
  19. transformers/dynamic_module_utils.py +25 -24
  20. transformers/feature_extraction_sequence_utils.py +23 -19
  21. transformers/feature_extraction_utils.py +33 -64
  22. transformers/file_utils.py +1 -0
  23. transformers/generation/__init__.py +1 -11
  24. transformers/generation/candidate_generator.py +33 -80
  25. transformers/generation/configuration_utils.py +133 -189
  26. transformers/generation/continuous_batching/__init__.py +1 -4
  27. transformers/generation/continuous_batching/cache.py +25 -83
  28. transformers/generation/continuous_batching/cache_manager.py +45 -155
  29. transformers/generation/continuous_batching/continuous_api.py +147 -270
  30. transformers/generation/continuous_batching/requests.py +3 -51
  31. transformers/generation/continuous_batching/scheduler.py +105 -160
  32. transformers/generation/logits_process.py +128 -0
  33. transformers/generation/stopping_criteria.py +1 -1
  34. transformers/generation/streamers.py +1 -0
  35. transformers/generation/utils.py +123 -122
  36. transformers/generation/watermarking.py +6 -8
  37. transformers/hf_argparser.py +13 -9
  38. transformers/hyperparameter_search.py +2 -1
  39. transformers/image_processing_base.py +23 -12
  40. transformers/image_processing_utils.py +15 -11
  41. transformers/image_processing_utils_fast.py +75 -85
  42. transformers/image_transforms.py +42 -73
  43. transformers/image_utils.py +32 -30
  44. transformers/initialization.py +0 -37
  45. transformers/integrations/__init__.py +2 -16
  46. transformers/integrations/accelerate.py +113 -58
  47. transformers/integrations/aqlm.py +66 -36
  48. transformers/integrations/awq.py +516 -45
  49. transformers/integrations/bitnet.py +105 -47
  50. transformers/integrations/bitsandbytes.py +202 -91
  51. transformers/integrations/deepspeed.py +4 -161
  52. transformers/integrations/eetq.py +82 -84
  53. transformers/integrations/executorch.py +1 -1
  54. transformers/integrations/fbgemm_fp8.py +145 -190
  55. transformers/integrations/finegrained_fp8.py +215 -249
  56. transformers/integrations/flash_attention.py +3 -3
  57. transformers/integrations/flex_attention.py +1 -1
  58. transformers/integrations/fp_quant.py +0 -90
  59. transformers/integrations/ggml.py +2 -11
  60. transformers/integrations/higgs.py +62 -37
  61. transformers/integrations/hub_kernels.py +8 -65
  62. transformers/integrations/integration_utils.py +3 -47
  63. transformers/integrations/mistral.py +0 -12
  64. transformers/integrations/mxfp4.py +80 -33
  65. transformers/integrations/peft.py +191 -483
  66. transformers/integrations/quanto.py +56 -77
  67. transformers/integrations/spqr.py +90 -42
  68. transformers/integrations/tensor_parallel.py +221 -167
  69. transformers/integrations/torchao.py +43 -35
  70. transformers/integrations/vptq.py +59 -40
  71. transformers/kernels/__init__.py +0 -0
  72. transformers/{models/pe_audio_video/processing_pe_audio_video.py → kernels/falcon_mamba/__init__.py} +3 -12
  73. transformers/kernels/falcon_mamba/selective_scan_with_ln_interface.py +529 -0
  74. transformers/loss/loss_utils.py +0 -2
  75. transformers/masking_utils.py +55 -51
  76. transformers/model_debugging_utils.py +5 -4
  77. transformers/modelcard.py +194 -15
  78. transformers/modeling_attn_mask_utils.py +19 -19
  79. transformers/modeling_flash_attention_utils.py +27 -27
  80. transformers/modeling_gguf_pytorch_utils.py +24 -79
  81. transformers/modeling_layers.py +22 -21
  82. transformers/modeling_outputs.py +253 -242
  83. transformers/modeling_rope_utils.py +117 -138
  84. transformers/modeling_utils.py +739 -850
  85. transformers/models/__init__.py +0 -27
  86. transformers/models/afmoe/configuration_afmoe.py +33 -40
  87. transformers/models/afmoe/modeling_afmoe.py +54 -42
  88. transformers/models/afmoe/modular_afmoe.py +33 -23
  89. transformers/models/aimv2/configuration_aimv2.py +10 -2
  90. transformers/models/aimv2/modeling_aimv2.py +42 -47
  91. transformers/models/aimv2/modular_aimv2.py +19 -17
  92. transformers/models/albert/configuration_albert.py +2 -8
  93. transformers/models/albert/modeling_albert.py +69 -70
  94. transformers/models/albert/tokenization_albert.py +14 -5
  95. transformers/models/align/configuration_align.py +6 -8
  96. transformers/models/align/modeling_align.py +89 -94
  97. transformers/models/align/processing_align.py +30 -2
  98. transformers/models/altclip/configuration_altclip.py +7 -4
  99. transformers/models/altclip/modeling_altclip.py +103 -114
  100. transformers/models/altclip/processing_altclip.py +15 -2
  101. transformers/models/apertus/__init__.py +1 -0
  102. transformers/models/apertus/configuration_apertus.py +28 -23
  103. transformers/models/apertus/modeling_apertus.py +40 -39
  104. transformers/models/apertus/modular_apertus.py +38 -37
  105. transformers/models/arcee/configuration_arcee.py +30 -25
  106. transformers/models/arcee/modeling_arcee.py +39 -36
  107. transformers/models/arcee/modular_arcee.py +23 -20
  108. transformers/models/aria/configuration_aria.py +44 -31
  109. transformers/models/aria/image_processing_aria.py +27 -25
  110. transformers/models/aria/modeling_aria.py +106 -110
  111. transformers/models/aria/modular_aria.py +127 -118
  112. transformers/models/aria/processing_aria.py +35 -28
  113. transformers/models/audio_spectrogram_transformer/configuration_audio_spectrogram_transformer.py +1 -0
  114. transformers/models/audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.py +6 -3
  115. transformers/models/audio_spectrogram_transformer/modeling_audio_spectrogram_transformer.py +8 -6
  116. transformers/models/audioflamingo3/__init__.py +1 -0
  117. transformers/models/audioflamingo3/configuration_audioflamingo3.py +1 -0
  118. transformers/models/audioflamingo3/modeling_audioflamingo3.py +49 -58
  119. transformers/models/audioflamingo3/modular_audioflamingo3.py +43 -53
  120. transformers/models/audioflamingo3/processing_audioflamingo3.py +30 -33
  121. transformers/models/auto/auto_factory.py +7 -6
  122. transformers/models/auto/configuration_auto.py +5 -66
  123. transformers/models/auto/feature_extraction_auto.py +10 -14
  124. transformers/models/auto/image_processing_auto.py +41 -32
  125. transformers/models/auto/modeling_auto.py +188 -46
  126. transformers/models/auto/processing_auto.py +11 -24
  127. transformers/models/auto/tokenization_auto.py +588 -171
  128. transformers/models/auto/video_processing_auto.py +10 -12
  129. transformers/models/autoformer/configuration_autoformer.py +7 -4
  130. transformers/models/autoformer/modeling_autoformer.py +101 -104
  131. transformers/models/aya_vision/configuration_aya_vision.py +1 -4
  132. transformers/models/aya_vision/modeling_aya_vision.py +102 -71
  133. transformers/models/aya_vision/modular_aya_vision.py +74 -46
  134. transformers/models/aya_vision/processing_aya_vision.py +53 -25
  135. transformers/models/bamba/configuration_bamba.py +39 -34
  136. transformers/models/bamba/modeling_bamba.py +86 -82
  137. transformers/models/bamba/modular_bamba.py +72 -70
  138. transformers/models/bark/configuration_bark.py +8 -6
  139. transformers/models/bark/generation_configuration_bark.py +5 -3
  140. transformers/models/bark/modeling_bark.py +57 -54
  141. transformers/models/bark/processing_bark.py +41 -19
  142. transformers/models/bart/configuration_bart.py +6 -9
  143. transformers/models/bart/modeling_bart.py +126 -135
  144. transformers/models/barthez/tokenization_barthez.py +11 -3
  145. transformers/models/bartpho/tokenization_bartpho.py +7 -6
  146. transformers/models/beit/configuration_beit.py +11 -0
  147. transformers/models/beit/image_processing_beit.py +56 -53
  148. transformers/models/beit/image_processing_beit_fast.py +12 -10
  149. transformers/models/beit/modeling_beit.py +60 -69
  150. transformers/models/bert/configuration_bert.py +2 -12
  151. transformers/models/bert/modeling_bert.py +122 -114
  152. transformers/models/bert/tokenization_bert.py +23 -8
  153. transformers/models/bert/tokenization_bert_legacy.py +5 -3
  154. transformers/models/bert_generation/configuration_bert_generation.py +2 -17
  155. transformers/models/bert_generation/modeling_bert_generation.py +49 -49
  156. transformers/models/bert_generation/tokenization_bert_generation.py +3 -2
  157. transformers/models/bert_japanese/tokenization_bert_japanese.py +6 -5
  158. transformers/models/bertweet/tokenization_bertweet.py +3 -1
  159. transformers/models/big_bird/configuration_big_bird.py +9 -12
  160. transformers/models/big_bird/modeling_big_bird.py +109 -116
  161. transformers/models/big_bird/tokenization_big_bird.py +43 -16
  162. transformers/models/bigbird_pegasus/configuration_bigbird_pegasus.py +9 -9
  163. transformers/models/bigbird_pegasus/modeling_bigbird_pegasus.py +117 -130
  164. transformers/models/biogpt/configuration_biogpt.py +2 -8
  165. transformers/models/biogpt/modeling_biogpt.py +76 -72
  166. transformers/models/biogpt/modular_biogpt.py +66 -62
  167. transformers/models/biogpt/tokenization_biogpt.py +5 -3
  168. transformers/models/bit/configuration_bit.py +1 -0
  169. transformers/models/bit/image_processing_bit.py +24 -21
  170. transformers/models/bit/image_processing_bit_fast.py +1 -0
  171. transformers/models/bit/modeling_bit.py +12 -25
  172. transformers/models/bitnet/configuration_bitnet.py +28 -23
  173. transformers/models/bitnet/modeling_bitnet.py +39 -36
  174. transformers/models/bitnet/modular_bitnet.py +6 -4
  175. transformers/models/blenderbot/configuration_blenderbot.py +5 -8
  176. transformers/models/blenderbot/modeling_blenderbot.py +96 -77
  177. transformers/models/blenderbot/tokenization_blenderbot.py +24 -18
  178. transformers/models/blenderbot_small/configuration_blenderbot_small.py +5 -8
  179. transformers/models/blenderbot_small/modeling_blenderbot_small.py +69 -79
  180. transformers/models/blenderbot_small/tokenization_blenderbot_small.py +3 -1
  181. transformers/models/blip/configuration_blip.py +10 -9
  182. transformers/models/blip/image_processing_blip.py +20 -17
  183. transformers/models/blip/image_processing_blip_fast.py +1 -0
  184. transformers/models/blip/modeling_blip.py +108 -117
  185. transformers/models/blip/modeling_blip_text.py +65 -73
  186. transformers/models/blip/processing_blip.py +36 -5
  187. transformers/models/blip_2/configuration_blip_2.py +2 -2
  188. transformers/models/blip_2/modeling_blip_2.py +118 -146
  189. transformers/models/blip_2/processing_blip_2.py +38 -8
  190. transformers/models/bloom/configuration_bloom.py +2 -5
  191. transformers/models/bloom/modeling_bloom.py +104 -77
  192. transformers/models/blt/configuration_blt.py +86 -94
  193. transformers/models/blt/modeling_blt.py +81 -238
  194. transformers/models/blt/modular_blt.py +65 -228
  195. transformers/models/bridgetower/configuration_bridgetower.py +2 -7
  196. transformers/models/bridgetower/image_processing_bridgetower.py +35 -34
  197. transformers/models/bridgetower/image_processing_bridgetower_fast.py +16 -13
  198. transformers/models/bridgetower/modeling_bridgetower.py +119 -141
  199. transformers/models/bridgetower/processing_bridgetower.py +16 -2
  200. transformers/models/bros/configuration_bros.py +18 -24
  201. transformers/models/bros/modeling_bros.py +80 -90
  202. transformers/models/bros/processing_bros.py +12 -2
  203. transformers/models/byt5/tokenization_byt5.py +6 -4
  204. transformers/models/camembert/configuration_camembert.py +2 -8
  205. transformers/models/camembert/modeling_camembert.py +195 -196
  206. transformers/models/camembert/modular_camembert.py +54 -51
  207. transformers/models/camembert/tokenization_camembert.py +13 -6
  208. transformers/models/canine/configuration_canine.py +2 -4
  209. transformers/models/canine/modeling_canine.py +75 -84
  210. transformers/models/canine/tokenization_canine.py +1 -2
  211. transformers/models/chameleon/configuration_chameleon.py +34 -29
  212. transformers/models/chameleon/image_processing_chameleon.py +24 -21
  213. transformers/models/chameleon/image_processing_chameleon_fast.py +6 -5
  214. transformers/models/chameleon/modeling_chameleon.py +93 -142
  215. transformers/models/chameleon/processing_chameleon.py +41 -16
  216. transformers/models/chinese_clip/configuration_chinese_clip.py +8 -10
  217. transformers/models/chinese_clip/image_processing_chinese_clip.py +24 -21
  218. transformers/models/chinese_clip/image_processing_chinese_clip_fast.py +1 -0
  219. transformers/models/chinese_clip/modeling_chinese_clip.py +92 -96
  220. transformers/models/chinese_clip/processing_chinese_clip.py +15 -2
  221. transformers/models/clap/configuration_clap.py +9 -4
  222. transformers/models/clap/feature_extraction_clap.py +12 -11
  223. transformers/models/clap/modeling_clap.py +123 -136
  224. transformers/models/clap/processing_clap.py +15 -2
  225. transformers/models/clip/configuration_clip.py +2 -4
  226. transformers/models/clip/image_processing_clip.py +24 -21
  227. transformers/models/clip/image_processing_clip_fast.py +1 -9
  228. transformers/models/clip/modeling_clip.py +65 -65
  229. transformers/models/clip/processing_clip.py +14 -2
  230. transformers/models/clip/tokenization_clip.py +46 -21
  231. transformers/models/clipseg/configuration_clipseg.py +2 -4
  232. transformers/models/clipseg/modeling_clipseg.py +109 -119
  233. transformers/models/clipseg/processing_clipseg.py +42 -19
  234. transformers/models/clvp/configuration_clvp.py +5 -15
  235. transformers/models/clvp/feature_extraction_clvp.py +10 -7
  236. transformers/models/clvp/modeling_clvp.py +146 -155
  237. transformers/models/clvp/number_normalizer.py +2 -1
  238. transformers/models/clvp/processing_clvp.py +20 -3
  239. transformers/models/clvp/tokenization_clvp.py +64 -1
  240. transformers/models/code_llama/tokenization_code_llama.py +44 -18
  241. transformers/models/codegen/configuration_codegen.py +4 -4
  242. transformers/models/codegen/modeling_codegen.py +53 -63
  243. transformers/models/codegen/tokenization_codegen.py +47 -17
  244. transformers/models/cohere/configuration_cohere.py +30 -25
  245. transformers/models/cohere/modeling_cohere.py +42 -40
  246. transformers/models/cohere/modular_cohere.py +29 -26
  247. transformers/models/cohere/tokenization_cohere.py +46 -15
  248. transformers/models/cohere2/configuration_cohere2.py +32 -31
  249. transformers/models/cohere2/modeling_cohere2.py +44 -42
  250. transformers/models/cohere2/modular_cohere2.py +54 -54
  251. transformers/models/cohere2_vision/image_processing_cohere2_vision_fast.py +14 -13
  252. transformers/models/cohere2_vision/modeling_cohere2_vision.py +58 -59
  253. transformers/models/cohere2_vision/modular_cohere2_vision.py +46 -45
  254. transformers/models/cohere2_vision/processing_cohere2_vision.py +36 -6
  255. transformers/models/colpali/configuration_colpali.py +1 -0
  256. transformers/models/colpali/modeling_colpali.py +16 -14
  257. transformers/models/colpali/modular_colpali.py +51 -11
  258. transformers/models/colpali/processing_colpali.py +52 -14
  259. transformers/models/colqwen2/modeling_colqwen2.py +28 -28
  260. transformers/models/colqwen2/modular_colqwen2.py +74 -37
  261. transformers/models/colqwen2/processing_colqwen2.py +52 -16
  262. transformers/models/conditional_detr/configuration_conditional_detr.py +2 -1
  263. transformers/models/conditional_detr/image_processing_conditional_detr.py +70 -67
  264. transformers/models/conditional_detr/image_processing_conditional_detr_fast.py +36 -36
  265. transformers/models/conditional_detr/modeling_conditional_detr.py +87 -99
  266. transformers/models/conditional_detr/modular_conditional_detr.py +3 -49
  267. transformers/models/convbert/configuration_convbert.py +8 -11
  268. transformers/models/convbert/modeling_convbert.py +87 -94
  269. transformers/models/convbert/tokenization_convbert.py +1 -0
  270. transformers/models/convnext/configuration_convnext.py +1 -0
  271. transformers/models/convnext/image_processing_convnext.py +23 -20
  272. transformers/models/convnext/image_processing_convnext_fast.py +21 -16
  273. transformers/models/convnext/modeling_convnext.py +12 -9
  274. transformers/models/convnextv2/configuration_convnextv2.py +1 -0
  275. transformers/models/convnextv2/modeling_convnextv2.py +12 -9
  276. transformers/models/cpm/tokenization_cpm.py +7 -6
  277. transformers/models/cpm/tokenization_cpm_fast.py +5 -3
  278. transformers/models/cpmant/configuration_cpmant.py +1 -4
  279. transformers/models/cpmant/modeling_cpmant.py +40 -38
  280. transformers/models/cpmant/tokenization_cpmant.py +3 -1
  281. transformers/models/csm/configuration_csm.py +66 -58
  282. transformers/models/csm/generation_csm.py +35 -31
  283. transformers/models/csm/modeling_csm.py +85 -85
  284. transformers/models/csm/modular_csm.py +58 -58
  285. transformers/models/csm/processing_csm.py +68 -25
  286. transformers/models/ctrl/configuration_ctrl.py +1 -16
  287. transformers/models/ctrl/modeling_ctrl.py +44 -54
  288. transformers/models/ctrl/tokenization_ctrl.py +1 -0
  289. transformers/models/cvt/configuration_cvt.py +1 -0
  290. transformers/models/cvt/modeling_cvt.py +16 -20
  291. transformers/models/cwm/__init__.py +1 -0
  292. transformers/models/cwm/configuration_cwm.py +12 -8
  293. transformers/models/cwm/modeling_cwm.py +39 -37
  294. transformers/models/cwm/modular_cwm.py +12 -10
  295. transformers/models/d_fine/configuration_d_fine.py +5 -7
  296. transformers/models/d_fine/modeling_d_fine.py +128 -138
  297. transformers/models/d_fine/modular_d_fine.py +18 -33
  298. transformers/models/dab_detr/configuration_dab_detr.py +3 -6
  299. transformers/models/dab_detr/modeling_dab_detr.py +75 -81
  300. transformers/models/dac/configuration_dac.py +1 -0
  301. transformers/models/dac/feature_extraction_dac.py +9 -6
  302. transformers/models/dac/modeling_dac.py +26 -24
  303. transformers/models/data2vec/configuration_data2vec_audio.py +2 -4
  304. transformers/models/data2vec/configuration_data2vec_text.py +3 -11
  305. transformers/models/data2vec/configuration_data2vec_vision.py +1 -0
  306. transformers/models/data2vec/modeling_data2vec_audio.py +56 -57
  307. transformers/models/data2vec/modeling_data2vec_text.py +93 -98
  308. transformers/models/data2vec/modeling_data2vec_vision.py +45 -49
  309. transformers/models/data2vec/modular_data2vec_audio.py +1 -6
  310. transformers/models/data2vec/modular_data2vec_text.py +54 -58
  311. transformers/models/dbrx/configuration_dbrx.py +22 -36
  312. transformers/models/dbrx/modeling_dbrx.py +45 -42
  313. transformers/models/dbrx/modular_dbrx.py +33 -31
  314. transformers/models/deberta/configuration_deberta.py +1 -6
  315. transformers/models/deberta/modeling_deberta.py +60 -64
  316. transformers/models/deberta/tokenization_deberta.py +21 -9
  317. transformers/models/deberta_v2/configuration_deberta_v2.py +1 -6
  318. transformers/models/deberta_v2/modeling_deberta_v2.py +65 -71
  319. transformers/models/deberta_v2/tokenization_deberta_v2.py +29 -11
  320. transformers/models/decision_transformer/configuration_decision_transformer.py +2 -3
  321. transformers/models/decision_transformer/modeling_decision_transformer.py +56 -60
  322. transformers/models/deepseek_v2/configuration_deepseek_v2.py +44 -39
  323. transformers/models/deepseek_v2/modeling_deepseek_v2.py +43 -43
  324. transformers/models/deepseek_v2/modular_deepseek_v2.py +49 -48
  325. transformers/models/deepseek_v3/configuration_deepseek_v3.py +45 -40
  326. transformers/models/deepseek_v3/modeling_deepseek_v3.py +42 -45
  327. transformers/models/deepseek_v3/modular_deepseek_v3.py +9 -14
  328. transformers/models/deepseek_vl/configuration_deepseek_vl.py +3 -2
  329. transformers/models/deepseek_vl/image_processing_deepseek_vl.py +26 -25
  330. transformers/models/deepseek_vl/image_processing_deepseek_vl_fast.py +10 -10
  331. transformers/models/deepseek_vl/modeling_deepseek_vl.py +48 -57
  332. transformers/models/deepseek_vl/modular_deepseek_vl.py +43 -14
  333. transformers/models/deepseek_vl/processing_deepseek_vl.py +41 -10
  334. transformers/models/deepseek_vl_hybrid/configuration_deepseek_vl_hybrid.py +5 -3
  335. transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py +35 -35
  336. transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py +24 -20
  337. transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py +61 -109
  338. transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py +118 -146
  339. transformers/models/deepseek_vl_hybrid/processing_deepseek_vl_hybrid.py +44 -12
  340. transformers/models/deformable_detr/configuration_deformable_detr.py +3 -2
  341. transformers/models/deformable_detr/image_processing_deformable_detr.py +61 -59
  342. transformers/models/deformable_detr/image_processing_deformable_detr_fast.py +28 -28
  343. transformers/models/deformable_detr/modeling_deformable_detr.py +82 -88
  344. transformers/models/deformable_detr/modular_deformable_detr.py +3 -1
  345. transformers/models/deit/configuration_deit.py +1 -0
  346. transformers/models/deit/image_processing_deit.py +21 -18
  347. transformers/models/deit/image_processing_deit_fast.py +1 -0
  348. transformers/models/deit/modeling_deit.py +22 -24
  349. transformers/models/depth_anything/configuration_depth_anything.py +4 -2
  350. transformers/models/depth_anything/modeling_depth_anything.py +10 -10
  351. transformers/models/depth_pro/configuration_depth_pro.py +1 -0
  352. transformers/models/depth_pro/image_processing_depth_pro.py +23 -22
  353. transformers/models/depth_pro/image_processing_depth_pro_fast.py +10 -8
  354. transformers/models/depth_pro/modeling_depth_pro.py +27 -31
  355. transformers/models/detr/configuration_detr.py +2 -1
  356. transformers/models/detr/image_processing_detr.py +66 -64
  357. transformers/models/detr/image_processing_detr_fast.py +34 -33
  358. transformers/models/detr/modeling_detr.py +79 -95
  359. transformers/models/dia/configuration_dia.py +15 -9
  360. transformers/models/dia/feature_extraction_dia.py +9 -6
  361. transformers/models/dia/generation_dia.py +50 -48
  362. transformers/models/dia/modeling_dia.py +69 -78
  363. transformers/models/dia/modular_dia.py +56 -64
  364. transformers/models/dia/processing_dia.py +29 -39
  365. transformers/models/dia/tokenization_dia.py +6 -3
  366. transformers/models/diffllama/configuration_diffllama.py +30 -25
  367. transformers/models/diffllama/modeling_diffllama.py +49 -46
  368. transformers/models/diffllama/modular_diffllama.py +19 -17
  369. transformers/models/dinat/configuration_dinat.py +1 -0
  370. transformers/models/dinat/modeling_dinat.py +44 -47
  371. transformers/models/dinov2/configuration_dinov2.py +1 -0
  372. transformers/models/dinov2/modeling_dinov2.py +15 -15
  373. transformers/models/dinov2_with_registers/configuration_dinov2_with_registers.py +1 -1
  374. transformers/models/dinov2_with_registers/modeling_dinov2_with_registers.py +15 -16
  375. transformers/models/dinov2_with_registers/modular_dinov2_with_registers.py +9 -9
  376. transformers/models/dinov3_convnext/configuration_dinov3_convnext.py +7 -4
  377. transformers/models/dinov3_convnext/modeling_dinov3_convnext.py +6 -3
  378. transformers/models/dinov3_vit/configuration_dinov3_vit.py +8 -5
  379. transformers/models/dinov3_vit/image_processing_dinov3_vit_fast.py +9 -7
  380. transformers/models/dinov3_vit/modeling_dinov3_vit.py +18 -19
  381. transformers/models/dinov3_vit/modular_dinov3_vit.py +15 -16
  382. transformers/models/distilbert/configuration_distilbert.py +2 -8
  383. transformers/models/distilbert/modeling_distilbert.py +55 -55
  384. transformers/models/distilbert/tokenization_distilbert.py +1 -13
  385. transformers/models/doge/__init__.py +1 -0
  386. transformers/models/doge/configuration_doge.py +32 -39
  387. transformers/models/doge/modeling_doge.py +49 -45
  388. transformers/models/doge/modular_doge.py +63 -71
  389. transformers/models/donut/configuration_donut_swin.py +1 -0
  390. transformers/models/donut/image_processing_donut.py +29 -26
  391. transformers/models/donut/image_processing_donut_fast.py +15 -9
  392. transformers/models/donut/modeling_donut_swin.py +58 -62
  393. transformers/models/donut/processing_donut.py +26 -5
  394. transformers/models/dots1/configuration_dots1.py +33 -41
  395. transformers/models/dots1/modeling_dots1.py +45 -54
  396. transformers/models/dots1/modular_dots1.py +4 -5
  397. transformers/models/dpr/configuration_dpr.py +2 -19
  398. transformers/models/dpr/modeling_dpr.py +39 -42
  399. transformers/models/dpr/tokenization_dpr.py +9 -19
  400. transformers/models/dpr/tokenization_dpr_fast.py +9 -7
  401. transformers/models/dpt/configuration_dpt.py +2 -1
  402. transformers/models/dpt/image_processing_dpt.py +66 -65
  403. transformers/models/dpt/image_processing_dpt_fast.py +20 -18
  404. transformers/models/dpt/modeling_dpt.py +30 -32
  405. transformers/models/dpt/modular_dpt.py +17 -15
  406. transformers/models/edgetam/configuration_edgetam.py +3 -2
  407. transformers/models/edgetam/modeling_edgetam.py +86 -86
  408. transformers/models/edgetam/modular_edgetam.py +26 -21
  409. transformers/models/edgetam_video/__init__.py +1 -0
  410. transformers/models/edgetam_video/configuration_edgetam_video.py +1 -0
  411. transformers/models/edgetam_video/modeling_edgetam_video.py +158 -169
  412. transformers/models/edgetam_video/modular_edgetam_video.py +37 -30
  413. transformers/models/efficientloftr/configuration_efficientloftr.py +5 -4
  414. transformers/models/efficientloftr/image_processing_efficientloftr.py +16 -14
  415. transformers/models/efficientloftr/image_processing_efficientloftr_fast.py +9 -9
  416. transformers/models/efficientloftr/modeling_efficientloftr.py +38 -59
  417. transformers/models/efficientloftr/modular_efficientloftr.py +3 -1
  418. transformers/models/efficientnet/configuration_efficientnet.py +1 -0
  419. transformers/models/efficientnet/image_processing_efficientnet.py +32 -28
  420. transformers/models/efficientnet/image_processing_efficientnet_fast.py +19 -17
  421. transformers/models/efficientnet/modeling_efficientnet.py +15 -19
  422. transformers/models/electra/configuration_electra.py +3 -13
  423. transformers/models/electra/modeling_electra.py +103 -108
  424. transformers/models/emu3/configuration_emu3.py +17 -13
  425. transformers/models/emu3/image_processing_emu3.py +39 -44
  426. transformers/models/emu3/modeling_emu3.py +108 -148
  427. transformers/models/emu3/modular_emu3.py +73 -115
  428. transformers/models/emu3/processing_emu3.py +43 -18
  429. transformers/models/encodec/configuration_encodec.py +4 -2
  430. transformers/models/encodec/feature_extraction_encodec.py +13 -10
  431. transformers/models/encodec/modeling_encodec.py +29 -39
  432. transformers/models/encoder_decoder/configuration_encoder_decoder.py +2 -12
  433. transformers/models/encoder_decoder/modeling_encoder_decoder.py +43 -37
  434. transformers/models/eomt/configuration_eomt.py +1 -0
  435. transformers/models/eomt/image_processing_eomt.py +56 -66
  436. transformers/models/eomt/image_processing_eomt_fast.py +33 -76
  437. transformers/models/eomt/modeling_eomt.py +18 -23
  438. transformers/models/eomt/modular_eomt.py +13 -18
  439. transformers/models/ernie/configuration_ernie.py +3 -24
  440. transformers/models/ernie/modeling_ernie.py +132 -127
  441. transformers/models/ernie/modular_ernie.py +103 -97
  442. transformers/models/ernie4_5/configuration_ernie4_5.py +27 -23
  443. transformers/models/ernie4_5/modeling_ernie4_5.py +38 -36
  444. transformers/models/ernie4_5/modular_ernie4_5.py +4 -3
  445. transformers/models/ernie4_5_moe/configuration_ernie4_5_moe.py +36 -32
  446. transformers/models/ernie4_5_moe/modeling_ernie4_5_moe.py +55 -56
  447. transformers/models/ernie4_5_moe/modular_ernie4_5_moe.py +46 -18
  448. transformers/models/esm/configuration_esm.py +15 -11
  449. transformers/models/esm/modeling_esm.py +34 -38
  450. transformers/models/esm/modeling_esmfold.py +49 -53
  451. transformers/models/esm/openfold_utils/chunk_utils.py +6 -6
  452. transformers/models/esm/openfold_utils/loss.py +2 -1
  453. transformers/models/esm/openfold_utils/protein.py +16 -15
  454. transformers/models/esm/openfold_utils/tensor_utils.py +6 -6
  455. transformers/models/esm/tokenization_esm.py +4 -2
  456. transformers/models/evolla/configuration_evolla.py +40 -50
  457. transformers/models/evolla/modeling_evolla.py +66 -71
  458. transformers/models/evolla/modular_evolla.py +47 -53
  459. transformers/models/evolla/processing_evolla.py +35 -23
  460. transformers/models/exaone4/configuration_exaone4.py +25 -23
  461. transformers/models/exaone4/modeling_exaone4.py +38 -35
  462. transformers/models/exaone4/modular_exaone4.py +46 -44
  463. transformers/models/falcon/configuration_falcon.py +26 -31
  464. transformers/models/falcon/modeling_falcon.py +80 -82
  465. transformers/models/falcon_h1/configuration_falcon_h1.py +51 -45
  466. transformers/models/falcon_h1/modeling_falcon_h1.py +82 -85
  467. transformers/models/falcon_h1/modular_falcon_h1.py +51 -56
  468. transformers/models/falcon_mamba/configuration_falcon_mamba.py +2 -1
  469. transformers/models/falcon_mamba/modeling_falcon_mamba.py +82 -75
  470. transformers/models/falcon_mamba/modular_falcon_mamba.py +45 -28
  471. transformers/models/fastspeech2_conformer/configuration_fastspeech2_conformer.py +6 -2
  472. transformers/models/fastspeech2_conformer/modeling_fastspeech2_conformer.py +60 -76
  473. transformers/models/fastspeech2_conformer/tokenization_fastspeech2_conformer.py +3 -2
  474. transformers/models/flaubert/configuration_flaubert.py +5 -10
  475. transformers/models/flaubert/modeling_flaubert.py +143 -145
  476. transformers/models/flaubert/tokenization_flaubert.py +5 -3
  477. transformers/models/flava/configuration_flava.py +6 -5
  478. transformers/models/flava/image_processing_flava.py +67 -66
  479. transformers/models/flava/image_processing_flava_fast.py +49 -46
  480. transformers/models/flava/modeling_flava.py +136 -153
  481. transformers/models/flava/processing_flava.py +12 -2
  482. transformers/models/flex_olmo/__init__.py +1 -0
  483. transformers/models/flex_olmo/configuration_flex_olmo.py +32 -28
  484. transformers/models/flex_olmo/modeling_flex_olmo.py +47 -47
  485. transformers/models/flex_olmo/modular_flex_olmo.py +44 -40
  486. transformers/models/florence2/configuration_florence2.py +1 -0
  487. transformers/models/florence2/modeling_florence2.py +69 -111
  488. transformers/models/florence2/modular_florence2.py +101 -104
  489. transformers/models/florence2/processing_florence2.py +47 -18
  490. transformers/models/fnet/configuration_fnet.py +2 -6
  491. transformers/models/fnet/modeling_fnet.py +80 -83
  492. transformers/models/fnet/tokenization_fnet.py +1 -0
  493. transformers/models/focalnet/configuration_focalnet.py +1 -0
  494. transformers/models/focalnet/modeling_focalnet.py +45 -51
  495. transformers/models/fsmt/configuration_fsmt.py +17 -12
  496. transformers/models/fsmt/modeling_fsmt.py +48 -49
  497. transformers/models/fsmt/tokenization_fsmt.py +5 -3
  498. transformers/models/funnel/configuration_funnel.py +1 -8
  499. transformers/models/funnel/modeling_funnel.py +93 -99
  500. transformers/models/funnel/tokenization_funnel.py +27 -17
  501. transformers/models/fuyu/configuration_fuyu.py +34 -28
  502. transformers/models/fuyu/image_processing_fuyu.py +31 -29
  503. transformers/models/fuyu/image_processing_fuyu_fast.py +17 -17
  504. transformers/models/fuyu/modeling_fuyu.py +53 -53
  505. transformers/models/fuyu/processing_fuyu.py +34 -23
  506. transformers/models/gemma/configuration_gemma.py +30 -25
  507. transformers/models/gemma/modeling_gemma.py +50 -46
  508. transformers/models/gemma/modular_gemma.py +47 -42
  509. transformers/models/gemma/tokenization_gemma.py +30 -10
  510. transformers/models/gemma2/configuration_gemma2.py +35 -30
  511. transformers/models/gemma2/modeling_gemma2.py +42 -39
  512. transformers/models/gemma2/modular_gemma2.py +66 -63
  513. transformers/models/gemma3/configuration_gemma3.py +44 -44
  514. transformers/models/gemma3/image_processing_gemma3.py +31 -29
  515. transformers/models/gemma3/image_processing_gemma3_fast.py +13 -11
  516. transformers/models/gemma3/modeling_gemma3.py +207 -159
  517. transformers/models/gemma3/modular_gemma3.py +204 -153
  518. transformers/models/gemma3/processing_gemma3.py +5 -5
  519. transformers/models/gemma3n/configuration_gemma3n.py +26 -36
  520. transformers/models/gemma3n/feature_extraction_gemma3n.py +11 -9
  521. transformers/models/gemma3n/modeling_gemma3n.py +356 -222
  522. transformers/models/gemma3n/modular_gemma3n.py +207 -230
  523. transformers/models/gemma3n/processing_gemma3n.py +26 -12
  524. transformers/models/git/configuration_git.py +8 -5
  525. transformers/models/git/modeling_git.py +204 -266
  526. transformers/models/git/processing_git.py +14 -2
  527. transformers/models/glm/configuration_glm.py +28 -24
  528. transformers/models/glm/modeling_glm.py +40 -37
  529. transformers/models/glm/modular_glm.py +7 -4
  530. transformers/models/glm4/configuration_glm4.py +28 -24
  531. transformers/models/glm4/modeling_glm4.py +42 -40
  532. transformers/models/glm4/modular_glm4.py +10 -8
  533. transformers/models/glm46v/configuration_glm46v.py +1 -0
  534. transformers/models/glm46v/image_processing_glm46v.py +40 -35
  535. transformers/models/glm46v/image_processing_glm46v_fast.py +9 -9
  536. transformers/models/glm46v/modeling_glm46v.py +90 -137
  537. transformers/models/glm46v/modular_glm46v.py +3 -4
  538. transformers/models/glm46v/processing_glm46v.py +41 -7
  539. transformers/models/glm46v/video_processing_glm46v.py +11 -9
  540. transformers/models/glm4_moe/configuration_glm4_moe.py +32 -40
  541. transformers/models/glm4_moe/modeling_glm4_moe.py +42 -45
  542. transformers/models/glm4_moe/modular_glm4_moe.py +34 -42
  543. transformers/models/glm4v/configuration_glm4v.py +20 -18
  544. transformers/models/glm4v/image_processing_glm4v.py +40 -34
  545. transformers/models/glm4v/image_processing_glm4v_fast.py +9 -8
  546. transformers/models/glm4v/modeling_glm4v.py +205 -254
  547. transformers/models/glm4v/modular_glm4v.py +224 -210
  548. transformers/models/glm4v/processing_glm4v.py +41 -7
  549. transformers/models/glm4v/video_processing_glm4v.py +11 -9
  550. transformers/models/glm4v_moe/configuration_glm4v_moe.py +125 -136
  551. transformers/models/glm4v_moe/modeling_glm4v_moe.py +368 -377
  552. transformers/models/glm4v_moe/modular_glm4v_moe.py +169 -83
  553. transformers/models/glpn/configuration_glpn.py +1 -0
  554. transformers/models/glpn/image_processing_glpn.py +12 -11
  555. transformers/models/glpn/image_processing_glpn_fast.py +13 -11
  556. transformers/models/glpn/modeling_glpn.py +14 -16
  557. transformers/models/got_ocr2/configuration_got_ocr2.py +12 -4
  558. transformers/models/got_ocr2/image_processing_got_ocr2.py +24 -22
  559. transformers/models/got_ocr2/image_processing_got_ocr2_fast.py +11 -9
  560. transformers/models/got_ocr2/modeling_got_ocr2.py +80 -77
  561. transformers/models/got_ocr2/modular_got_ocr2.py +51 -54
  562. transformers/models/got_ocr2/processing_got_ocr2.py +63 -42
  563. transformers/models/gpt2/configuration_gpt2.py +2 -13
  564. transformers/models/gpt2/modeling_gpt2.py +115 -120
  565. transformers/models/gpt2/tokenization_gpt2.py +46 -15
  566. transformers/models/gpt_bigcode/configuration_gpt_bigcode.py +2 -5
  567. transformers/models/gpt_bigcode/modeling_gpt_bigcode.py +89 -79
  568. transformers/models/gpt_neo/configuration_gpt_neo.py +2 -9
  569. transformers/models/gpt_neo/modeling_gpt_neo.py +67 -83
  570. transformers/models/gpt_neox/configuration_gpt_neox.py +25 -25
  571. transformers/models/gpt_neox/modeling_gpt_neox.py +75 -76
  572. transformers/models/gpt_neox/modular_gpt_neox.py +66 -67
  573. transformers/models/gpt_neox/tokenization_gpt_neox.py +51 -9
  574. transformers/models/gpt_neox_japanese/configuration_gpt_neox_japanese.py +19 -24
  575. transformers/models/gpt_neox_japanese/modeling_gpt_neox_japanese.py +47 -46
  576. transformers/models/gpt_neox_japanese/tokenization_gpt_neox_japanese.py +3 -1
  577. transformers/models/gpt_oss/configuration_gpt_oss.py +28 -46
  578. transformers/models/gpt_oss/modeling_gpt_oss.py +121 -83
  579. transformers/models/gpt_oss/modular_gpt_oss.py +103 -64
  580. transformers/models/gpt_sw3/tokenization_gpt_sw3.py +4 -4
  581. transformers/models/gptj/configuration_gptj.py +4 -4
  582. transformers/models/gptj/modeling_gptj.py +87 -101
  583. transformers/models/granite/configuration_granite.py +33 -28
  584. transformers/models/granite/modeling_granite.py +46 -44
  585. transformers/models/granite/modular_granite.py +31 -29
  586. transformers/models/granite_speech/configuration_granite_speech.py +1 -0
  587. transformers/models/granite_speech/feature_extraction_granite_speech.py +3 -1
  588. transformers/models/granite_speech/modeling_granite_speech.py +52 -82
  589. transformers/models/granite_speech/processing_granite_speech.py +4 -11
  590. transformers/models/granitemoe/configuration_granitemoe.py +36 -31
  591. transformers/models/granitemoe/modeling_granitemoe.py +46 -41
  592. transformers/models/granitemoe/modular_granitemoe.py +27 -22
  593. transformers/models/granitemoehybrid/__init__.py +1 -0
  594. transformers/models/granitemoehybrid/configuration_granitemoehybrid.py +47 -46
  595. transformers/models/granitemoehybrid/modeling_granitemoehybrid.py +93 -97
  596. transformers/models/granitemoehybrid/modular_granitemoehybrid.py +21 -54
  597. transformers/models/granitemoeshared/configuration_granitemoeshared.py +37 -33
  598. transformers/models/granitemoeshared/modeling_granitemoeshared.py +61 -54
  599. transformers/models/granitemoeshared/modular_granitemoeshared.py +21 -19
  600. transformers/models/grounding_dino/configuration_grounding_dino.py +4 -6
  601. transformers/models/grounding_dino/image_processing_grounding_dino.py +62 -60
  602. transformers/models/grounding_dino/image_processing_grounding_dino_fast.py +29 -28
  603. transformers/models/grounding_dino/modeling_grounding_dino.py +140 -155
  604. transformers/models/grounding_dino/modular_grounding_dino.py +3 -2
  605. transformers/models/grounding_dino/processing_grounding_dino.py +38 -10
  606. transformers/models/groupvit/configuration_groupvit.py +2 -4
  607. transformers/models/groupvit/modeling_groupvit.py +93 -107
  608. transformers/models/helium/configuration_helium.py +29 -25
  609. transformers/models/helium/modeling_helium.py +40 -38
  610. transformers/models/helium/modular_helium.py +7 -3
  611. transformers/models/herbert/tokenization_herbert.py +28 -10
  612. transformers/models/hgnet_v2/configuration_hgnet_v2.py +1 -0
  613. transformers/models/hgnet_v2/modeling_hgnet_v2.py +10 -24
  614. transformers/models/hgnet_v2/modular_hgnet_v2.py +10 -24
  615. transformers/models/hiera/configuration_hiera.py +1 -0
  616. transformers/models/hiera/modeling_hiera.py +66 -72
  617. transformers/models/hubert/configuration_hubert.py +2 -4
  618. transformers/models/hubert/modeling_hubert.py +37 -42
  619. transformers/models/hubert/modular_hubert.py +11 -13
  620. transformers/models/hunyuan_v1_dense/configuration_hunyuan_v1_dense.py +31 -26
  621. transformers/models/hunyuan_v1_dense/modeling_hunyuan_v1_dense.py +38 -35
  622. transformers/models/hunyuan_v1_dense/modular_hunyuan_v1_dense.py +6 -4
  623. transformers/models/hunyuan_v1_moe/__init__.py +1 -1
  624. transformers/models/hunyuan_v1_moe/configuration_hunyuan_v1_moe.py +36 -31
  625. transformers/models/hunyuan_v1_moe/modeling_hunyuan_v1_moe.py +42 -47
  626. transformers/models/hunyuan_v1_moe/modular_hunyuan_v1_moe.py +9 -9
  627. transformers/models/ibert/configuration_ibert.py +2 -4
  628. transformers/models/ibert/modeling_ibert.py +62 -82
  629. transformers/models/ibert/quant_modules.py +1 -0
  630. transformers/models/idefics/configuration_idefics.py +8 -5
  631. transformers/models/idefics/image_processing_idefics.py +15 -13
  632. transformers/models/idefics/modeling_idefics.py +82 -75
  633. transformers/models/idefics/perceiver.py +3 -1
  634. transformers/models/idefics/processing_idefics.py +48 -32
  635. transformers/models/idefics/vision.py +25 -24
  636. transformers/models/idefics2/configuration_idefics2.py +3 -1
  637. transformers/models/idefics2/image_processing_idefics2.py +32 -31
  638. transformers/models/idefics2/image_processing_idefics2_fast.py +8 -8
  639. transformers/models/idefics2/modeling_idefics2.py +101 -127
  640. transformers/models/idefics2/processing_idefics2.py +68 -10
  641. transformers/models/idefics3/configuration_idefics3.py +4 -1
  642. transformers/models/idefics3/image_processing_idefics3.py +43 -42
  643. transformers/models/idefics3/image_processing_idefics3_fast.py +15 -40
  644. transformers/models/idefics3/modeling_idefics3.py +90 -115
  645. transformers/models/idefics3/processing_idefics3.py +69 -15
  646. transformers/models/ijepa/configuration_ijepa.py +1 -0
  647. transformers/models/ijepa/modeling_ijepa.py +11 -10
  648. transformers/models/ijepa/modular_ijepa.py +7 -5
  649. transformers/models/imagegpt/configuration_imagegpt.py +2 -9
  650. transformers/models/imagegpt/image_processing_imagegpt.py +18 -17
  651. transformers/models/imagegpt/image_processing_imagegpt_fast.py +16 -11
  652. transformers/models/imagegpt/modeling_imagegpt.py +65 -76
  653. transformers/models/informer/configuration_informer.py +9 -6
  654. transformers/models/informer/modeling_informer.py +86 -88
  655. transformers/models/informer/modular_informer.py +16 -14
  656. transformers/models/instructblip/configuration_instructblip.py +2 -2
  657. transformers/models/instructblip/modeling_instructblip.py +63 -103
  658. transformers/models/instructblip/processing_instructblip.py +36 -10
  659. transformers/models/instructblipvideo/configuration_instructblipvideo.py +2 -2
  660. transformers/models/instructblipvideo/modeling_instructblipvideo.py +139 -157
  661. transformers/models/instructblipvideo/modular_instructblipvideo.py +64 -73
  662. transformers/models/instructblipvideo/processing_instructblipvideo.py +33 -14
  663. transformers/models/instructblipvideo/video_processing_instructblipvideo.py +8 -6
  664. transformers/models/internvl/configuration_internvl.py +1 -0
  665. transformers/models/internvl/modeling_internvl.py +106 -85
  666. transformers/models/internvl/modular_internvl.py +67 -47
  667. transformers/models/internvl/processing_internvl.py +45 -12
  668. transformers/models/internvl/video_processing_internvl.py +12 -10
  669. transformers/models/jamba/configuration_jamba.py +8 -5
  670. transformers/models/jamba/modeling_jamba.py +66 -68
  671. transformers/models/jamba/modular_jamba.py +55 -54
  672. transformers/models/janus/configuration_janus.py +1 -0
  673. transformers/models/janus/image_processing_janus.py +37 -35
  674. transformers/models/janus/image_processing_janus_fast.py +20 -18
  675. transformers/models/janus/modeling_janus.py +191 -115
  676. transformers/models/janus/modular_janus.py +84 -133
  677. transformers/models/janus/processing_janus.py +43 -17
  678. transformers/models/jetmoe/configuration_jetmoe.py +26 -24
  679. transformers/models/jetmoe/modeling_jetmoe.py +46 -43
  680. transformers/models/jetmoe/modular_jetmoe.py +33 -31
  681. transformers/models/kosmos2/configuration_kosmos2.py +9 -10
  682. transformers/models/kosmos2/modeling_kosmos2.py +173 -208
  683. transformers/models/kosmos2/processing_kosmos2.py +55 -40
  684. transformers/models/kosmos2_5/__init__.py +1 -0
  685. transformers/models/kosmos2_5/configuration_kosmos2_5.py +9 -8
  686. transformers/models/kosmos2_5/image_processing_kosmos2_5.py +12 -10
  687. transformers/models/kosmos2_5/image_processing_kosmos2_5_fast.py +13 -4
  688. transformers/models/kosmos2_5/modeling_kosmos2_5.py +118 -132
  689. transformers/models/kosmos2_5/processing_kosmos2_5.py +29 -8
  690. transformers/models/kyutai_speech_to_text/configuration_kyutai_speech_to_text.py +28 -31
  691. transformers/models/kyutai_speech_to_text/feature_extraction_kyutai_speech_to_text.py +14 -12
  692. transformers/models/kyutai_speech_to_text/modeling_kyutai_speech_to_text.py +100 -110
  693. transformers/models/kyutai_speech_to_text/modular_kyutai_speech_to_text.py +22 -28
  694. transformers/models/kyutai_speech_to_text/processing_kyutai_speech_to_text.py +8 -2
  695. transformers/models/layoutlm/configuration_layoutlm.py +2 -14
  696. transformers/models/layoutlm/modeling_layoutlm.py +72 -77
  697. transformers/models/layoutlmv2/configuration_layoutlmv2.py +17 -14
  698. transformers/models/layoutlmv2/image_processing_layoutlmv2.py +21 -18
  699. transformers/models/layoutlmv2/image_processing_layoutlmv2_fast.py +9 -7
  700. transformers/models/layoutlmv2/modeling_layoutlmv2.py +50 -64
  701. transformers/models/layoutlmv2/processing_layoutlmv2.py +44 -14
  702. transformers/models/layoutlmv2/tokenization_layoutlmv2.py +126 -73
  703. transformers/models/layoutlmv3/configuration_layoutlmv3.py +19 -16
  704. transformers/models/layoutlmv3/image_processing_layoutlmv3.py +26 -24
  705. transformers/models/layoutlmv3/image_processing_layoutlmv3_fast.py +11 -9
  706. transformers/models/layoutlmv3/modeling_layoutlmv3.py +56 -82
  707. transformers/models/layoutlmv3/processing_layoutlmv3.py +46 -14
  708. transformers/models/layoutlmv3/tokenization_layoutlmv3.py +134 -74
  709. transformers/models/layoutxlm/configuration_layoutxlm.py +17 -14
  710. transformers/models/layoutxlm/modular_layoutxlm.py +1 -0
  711. transformers/models/layoutxlm/processing_layoutxlm.py +44 -14
  712. transformers/models/layoutxlm/tokenization_layoutxlm.py +113 -77
  713. transformers/models/led/configuration_led.py +12 -8
  714. transformers/models/led/modeling_led.py +266 -124
  715. transformers/models/levit/configuration_levit.py +1 -0
  716. transformers/models/levit/image_processing_levit.py +21 -19
  717. transformers/models/levit/image_processing_levit_fast.py +5 -4
  718. transformers/models/levit/modeling_levit.py +19 -38
  719. transformers/models/lfm2/configuration_lfm2.py +30 -27
  720. transformers/models/lfm2/modeling_lfm2.py +50 -47
  721. transformers/models/lfm2/modular_lfm2.py +30 -29
  722. transformers/models/lfm2_moe/__init__.py +1 -0
  723. transformers/models/lfm2_moe/configuration_lfm2_moe.py +9 -6
  724. transformers/models/lfm2_moe/modeling_lfm2_moe.py +53 -61
  725. transformers/models/lfm2_moe/modular_lfm2_moe.py +37 -13
  726. transformers/models/lfm2_vl/configuration_lfm2_vl.py +1 -4
  727. transformers/models/lfm2_vl/image_processing_lfm2_vl_fast.py +12 -41
  728. transformers/models/lfm2_vl/modeling_lfm2_vl.py +66 -84
  729. transformers/models/lfm2_vl/modular_lfm2_vl.py +56 -70
  730. transformers/models/lfm2_vl/processing_lfm2_vl.py +76 -96
  731. transformers/models/lightglue/image_processing_lightglue.py +15 -16
  732. transformers/models/lightglue/image_processing_lightglue_fast.py +9 -9
  733. transformers/models/lightglue/modeling_lightglue.py +31 -31
  734. transformers/models/lightglue/modular_lightglue.py +28 -29
  735. transformers/models/lilt/configuration_lilt.py +2 -6
  736. transformers/models/lilt/modeling_lilt.py +70 -76
  737. transformers/models/llama/configuration_llama.py +31 -26
  738. transformers/models/llama/modeling_llama.py +39 -36
  739. transformers/models/llama/tokenization_llama.py +44 -14
  740. transformers/models/llama4/configuration_llama4.py +30 -27
  741. transformers/models/llama4/image_processing_llama4_fast.py +14 -12
  742. transformers/models/llama4/modeling_llama4.py +113 -120
  743. transformers/models/llama4/processing_llama4.py +57 -33
  744. transformers/models/llava/configuration_llava.py +1 -10
  745. transformers/models/llava/image_processing_llava.py +28 -25
  746. transformers/models/llava/image_processing_llava_fast.py +11 -9
  747. transformers/models/llava/modeling_llava.py +109 -85
  748. transformers/models/llava/processing_llava.py +51 -18
  749. transformers/models/llava_next/configuration_llava_next.py +2 -2
  750. transformers/models/llava_next/image_processing_llava_next.py +45 -43
  751. transformers/models/llava_next/image_processing_llava_next_fast.py +13 -11
  752. transformers/models/llava_next/modeling_llava_next.py +107 -110
  753. transformers/models/llava_next/processing_llava_next.py +47 -18
  754. transformers/models/llava_next_video/configuration_llava_next_video.py +7 -4
  755. transformers/models/llava_next_video/modeling_llava_next_video.py +158 -175
  756. transformers/models/llava_next_video/modular_llava_next_video.py +150 -155
  757. transformers/models/llava_next_video/processing_llava_next_video.py +63 -21
  758. transformers/models/llava_next_video/video_processing_llava_next_video.py +1 -0
  759. transformers/models/llava_onevision/configuration_llava_onevision.py +7 -4
  760. transformers/models/llava_onevision/image_processing_llava_onevision.py +42 -40
  761. transformers/models/llava_onevision/image_processing_llava_onevision_fast.py +15 -14
  762. transformers/models/llava_onevision/modeling_llava_onevision.py +169 -177
  763. transformers/models/llava_onevision/modular_llava_onevision.py +156 -163
  764. transformers/models/llava_onevision/processing_llava_onevision.py +53 -21
  765. transformers/models/llava_onevision/video_processing_llava_onevision.py +1 -0
  766. transformers/models/longcat_flash/__init__.py +1 -0
  767. transformers/models/longcat_flash/configuration_longcat_flash.py +42 -37
  768. transformers/models/longcat_flash/modeling_longcat_flash.py +36 -36
  769. transformers/models/longcat_flash/modular_longcat_flash.py +21 -21
  770. transformers/models/longformer/configuration_longformer.py +5 -5
  771. transformers/models/longformer/modeling_longformer.py +101 -105
  772. transformers/models/longt5/configuration_longt5.py +7 -9
  773. transformers/models/longt5/modeling_longt5.py +49 -49
  774. transformers/models/luke/configuration_luke.py +2 -8
  775. transformers/models/luke/modeling_luke.py +181 -188
  776. transformers/models/luke/tokenization_luke.py +140 -107
  777. transformers/models/lxmert/configuration_lxmert.py +1 -16
  778. transformers/models/lxmert/modeling_lxmert.py +74 -65
  779. transformers/models/m2m_100/configuration_m2m_100.py +9 -7
  780. transformers/models/m2m_100/modeling_m2m_100.py +71 -83
  781. transformers/models/m2m_100/tokenization_m2m_100.py +8 -8
  782. transformers/models/mamba/configuration_mamba.py +2 -1
  783. transformers/models/mamba/modeling_mamba.py +66 -58
  784. transformers/models/mamba2/configuration_mamba2.py +8 -5
  785. transformers/models/mamba2/modeling_mamba2.py +69 -68
  786. transformers/models/marian/configuration_marian.py +5 -10
  787. transformers/models/marian/modeling_marian.py +87 -93
  788. transformers/models/marian/tokenization_marian.py +6 -6
  789. transformers/models/markuplm/configuration_markuplm.py +7 -4
  790. transformers/models/markuplm/feature_extraction_markuplm.py +2 -1
  791. transformers/models/markuplm/modeling_markuplm.py +70 -69
  792. transformers/models/markuplm/processing_markuplm.py +38 -31
  793. transformers/models/markuplm/tokenization_markuplm.py +136 -93
  794. transformers/models/mask2former/configuration_mask2former.py +8 -5
  795. transformers/models/mask2former/image_processing_mask2former.py +85 -84
  796. transformers/models/mask2former/image_processing_mask2former_fast.py +40 -37
  797. transformers/models/mask2former/modeling_mask2former.py +103 -118
  798. transformers/models/mask2former/modular_mask2former.py +8 -6
  799. transformers/models/maskformer/configuration_maskformer.py +9 -6
  800. transformers/models/maskformer/configuration_maskformer_swin.py +1 -0
  801. transformers/models/maskformer/image_processing_maskformer.py +85 -84
  802. transformers/models/maskformer/image_processing_maskformer_fast.py +40 -36
  803. transformers/models/maskformer/modeling_maskformer.py +65 -79
  804. transformers/models/maskformer/modeling_maskformer_swin.py +32 -36
  805. transformers/models/mbart/configuration_mbart.py +4 -9
  806. transformers/models/mbart/modeling_mbart.py +116 -131
  807. transformers/models/mbart/tokenization_mbart.py +54 -11
  808. transformers/models/mbart50/tokenization_mbart50.py +13 -8
  809. transformers/models/megatron_bert/configuration_megatron_bert.py +3 -13
  810. transformers/models/megatron_bert/modeling_megatron_bert.py +150 -148
  811. transformers/models/metaclip_2/configuration_metaclip_2.py +1 -4
  812. transformers/models/metaclip_2/modeling_metaclip_2.py +84 -91
  813. transformers/models/metaclip_2/modular_metaclip_2.py +45 -61
  814. transformers/models/mgp_str/configuration_mgp_str.py +1 -0
  815. transformers/models/mgp_str/modeling_mgp_str.py +18 -20
  816. transformers/models/mgp_str/processing_mgp_str.py +20 -3
  817. transformers/models/mgp_str/tokenization_mgp_str.py +3 -1
  818. transformers/models/mimi/configuration_mimi.py +40 -42
  819. transformers/models/mimi/modeling_mimi.py +113 -142
  820. transformers/models/minimax/__init__.py +1 -0
  821. transformers/models/minimax/configuration_minimax.py +43 -37
  822. transformers/models/minimax/modeling_minimax.py +51 -61
  823. transformers/models/minimax/modular_minimax.py +62 -68
  824. transformers/models/ministral/configuration_ministral.py +29 -25
  825. transformers/models/ministral/modeling_ministral.py +38 -36
  826. transformers/models/ministral/modular_ministral.py +37 -32
  827. transformers/models/ministral3/configuration_ministral3.py +27 -24
  828. transformers/models/ministral3/modeling_ministral3.py +37 -36
  829. transformers/models/ministral3/modular_ministral3.py +5 -4
  830. transformers/models/mistral/configuration_mistral.py +29 -24
  831. transformers/models/mistral/modeling_mistral.py +37 -36
  832. transformers/models/mistral/modular_mistral.py +12 -11
  833. transformers/models/mistral3/configuration_mistral3.py +1 -4
  834. transformers/models/mistral3/modeling_mistral3.py +86 -89
  835. transformers/models/mistral3/modular_mistral3.py +68 -69
  836. transformers/models/mixtral/configuration_mixtral.py +34 -29
  837. transformers/models/mixtral/modeling_mixtral.py +45 -50
  838. transformers/models/mixtral/modular_mixtral.py +31 -32
  839. transformers/models/mlcd/configuration_mlcd.py +1 -0
  840. transformers/models/mlcd/modeling_mlcd.py +14 -20
  841. transformers/models/mlcd/modular_mlcd.py +13 -17
  842. transformers/models/mllama/configuration_mllama.py +15 -10
  843. transformers/models/mllama/image_processing_mllama.py +25 -23
  844. transformers/models/mllama/image_processing_mllama_fast.py +11 -11
  845. transformers/models/mllama/modeling_mllama.py +94 -105
  846. transformers/models/mllama/processing_mllama.py +55 -6
  847. transformers/models/mluke/tokenization_mluke.py +107 -101
  848. transformers/models/mm_grounding_dino/configuration_mm_grounding_dino.py +3 -5
  849. transformers/models/mm_grounding_dino/modeling_mm_grounding_dino.py +140 -155
  850. transformers/models/mm_grounding_dino/modular_mm_grounding_dino.py +3 -5
  851. transformers/models/mobilebert/configuration_mobilebert.py +2 -4
  852. transformers/models/mobilebert/modeling_mobilebert.py +85 -77
  853. transformers/models/mobilebert/tokenization_mobilebert.py +1 -0
  854. transformers/models/mobilenet_v1/configuration_mobilenet_v1.py +1 -0
  855. transformers/models/mobilenet_v1/image_processing_mobilenet_v1.py +23 -20
  856. transformers/models/mobilenet_v1/image_processing_mobilenet_v1_fast.py +1 -0
  857. transformers/models/mobilenet_v1/modeling_mobilenet_v1.py +16 -15
  858. transformers/models/mobilenet_v2/configuration_mobilenet_v2.py +1 -0
  859. transformers/models/mobilenet_v2/image_processing_mobilenet_v2.py +51 -48
  860. transformers/models/mobilenet_v2/image_processing_mobilenet_v2_fast.py +15 -13
  861. transformers/models/mobilenet_v2/modeling_mobilenet_v2.py +22 -24
  862. transformers/models/mobilevit/configuration_mobilevit.py +1 -0
  863. transformers/models/mobilevit/image_processing_mobilevit.py +49 -46
  864. transformers/models/mobilevit/image_processing_mobilevit_fast.py +14 -12
  865. transformers/models/mobilevit/modeling_mobilevit.py +21 -28
  866. transformers/models/mobilevitv2/configuration_mobilevitv2.py +1 -0
  867. transformers/models/mobilevitv2/modeling_mobilevitv2.py +22 -28
  868. transformers/models/modernbert/configuration_modernbert.py +42 -44
  869. transformers/models/modernbert/modeling_modernbert.py +133 -145
  870. transformers/models/modernbert/modular_modernbert.py +170 -186
  871. transformers/models/modernbert_decoder/configuration_modernbert_decoder.py +40 -40
  872. transformers/models/modernbert_decoder/modeling_modernbert_decoder.py +57 -62
  873. transformers/models/modernbert_decoder/modular_modernbert_decoder.py +86 -94
  874. transformers/models/moonshine/configuration_moonshine.py +31 -34
  875. transformers/models/moonshine/modeling_moonshine.py +71 -71
  876. transformers/models/moonshine/modular_moonshine.py +83 -88
  877. transformers/models/moshi/configuration_moshi.py +23 -46
  878. transformers/models/moshi/modeling_moshi.py +187 -157
  879. transformers/models/mpnet/configuration_mpnet.py +2 -6
  880. transformers/models/mpnet/modeling_mpnet.py +57 -62
  881. transformers/models/mpnet/tokenization_mpnet.py +15 -4
  882. transformers/models/mpt/configuration_mpt.py +9 -5
  883. transformers/models/mpt/modeling_mpt.py +60 -60
  884. transformers/models/mra/configuration_mra.py +2 -8
  885. transformers/models/mra/modeling_mra.py +57 -64
  886. transformers/models/mt5/configuration_mt5.py +8 -10
  887. transformers/models/mt5/modeling_mt5.py +95 -87
  888. transformers/models/musicgen/configuration_musicgen.py +8 -12
  889. transformers/models/musicgen/modeling_musicgen.py +122 -118
  890. transformers/models/musicgen/processing_musicgen.py +21 -3
  891. transformers/models/musicgen_melody/configuration_musicgen_melody.py +8 -15
  892. transformers/models/musicgen_melody/feature_extraction_musicgen_melody.py +9 -8
  893. transformers/models/musicgen_melody/modeling_musicgen_melody.py +123 -117
  894. transformers/models/musicgen_melody/processing_musicgen_melody.py +22 -3
  895. transformers/models/mvp/configuration_mvp.py +5 -8
  896. transformers/models/mvp/modeling_mvp.py +123 -135
  897. transformers/models/myt5/tokenization_myt5.py +10 -8
  898. transformers/models/nanochat/configuration_nanochat.py +8 -5
  899. transformers/models/nanochat/modeling_nanochat.py +40 -37
  900. transformers/models/nanochat/modular_nanochat.py +14 -12
  901. transformers/models/nemotron/configuration_nemotron.py +30 -25
  902. transformers/models/nemotron/modeling_nemotron.py +57 -56
  903. transformers/models/nllb/tokenization_nllb.py +28 -12
  904. transformers/models/nllb_moe/configuration_nllb_moe.py +9 -7
  905. transformers/models/nllb_moe/modeling_nllb_moe.py +69 -77
  906. transformers/models/nougat/image_processing_nougat.py +32 -29
  907. transformers/models/nougat/image_processing_nougat_fast.py +14 -12
  908. transformers/models/nougat/processing_nougat.py +39 -37
  909. transformers/models/nougat/tokenization_nougat.py +73 -18
  910. transformers/models/nystromformer/configuration_nystromformer.py +2 -8
  911. transformers/models/nystromformer/modeling_nystromformer.py +63 -74
  912. transformers/models/olmo/configuration_olmo.py +28 -23
  913. transformers/models/olmo/modeling_olmo.py +39 -36
  914. transformers/models/olmo/modular_olmo.py +11 -7
  915. transformers/models/olmo2/configuration_olmo2.py +28 -23
  916. transformers/models/olmo2/modeling_olmo2.py +41 -37
  917. transformers/models/olmo2/modular_olmo2.py +32 -29
  918. transformers/models/olmo3/__init__.py +1 -0
  919. transformers/models/olmo3/configuration_olmo3.py +30 -26
  920. transformers/models/olmo3/modeling_olmo3.py +39 -36
  921. transformers/models/olmo3/modular_olmo3.py +40 -37
  922. transformers/models/olmoe/configuration_olmoe.py +33 -29
  923. transformers/models/olmoe/modeling_olmoe.py +46 -52
  924. transformers/models/olmoe/modular_olmoe.py +15 -16
  925. transformers/models/omdet_turbo/configuration_omdet_turbo.py +4 -2
  926. transformers/models/omdet_turbo/modeling_omdet_turbo.py +47 -53
  927. transformers/models/omdet_turbo/processing_omdet_turbo.py +67 -19
  928. transformers/models/oneformer/configuration_oneformer.py +8 -5
  929. transformers/models/oneformer/image_processing_oneformer.py +84 -83
  930. transformers/models/oneformer/image_processing_oneformer_fast.py +42 -41
  931. transformers/models/oneformer/modeling_oneformer.py +171 -147
  932. transformers/models/oneformer/processing_oneformer.py +43 -28
  933. transformers/models/openai/configuration_openai.py +1 -16
  934. transformers/models/openai/modeling_openai.py +51 -65
  935. transformers/models/openai/tokenization_openai.py +47 -8
  936. transformers/models/opt/configuration_opt.py +7 -6
  937. transformers/models/opt/modeling_opt.py +76 -78
  938. transformers/models/ovis2/__init__.py +1 -0
  939. transformers/models/ovis2/configuration_ovis2.py +1 -0
  940. transformers/models/ovis2/image_processing_ovis2.py +24 -22
  941. transformers/models/ovis2/image_processing_ovis2_fast.py +11 -9
  942. transformers/models/ovis2/modeling_ovis2.py +142 -111
  943. transformers/models/ovis2/modular_ovis2.py +45 -90
  944. transformers/models/ovis2/processing_ovis2.py +40 -12
  945. transformers/models/owlv2/configuration_owlv2.py +2 -4
  946. transformers/models/owlv2/image_processing_owlv2.py +21 -20
  947. transformers/models/owlv2/image_processing_owlv2_fast.py +15 -12
  948. transformers/models/owlv2/modeling_owlv2.py +117 -133
  949. transformers/models/owlv2/modular_owlv2.py +14 -11
  950. transformers/models/owlv2/processing_owlv2.py +49 -20
  951. transformers/models/owlvit/configuration_owlvit.py +2 -4
  952. transformers/models/owlvit/image_processing_owlvit.py +22 -21
  953. transformers/models/owlvit/image_processing_owlvit_fast.py +3 -2
  954. transformers/models/owlvit/modeling_owlvit.py +116 -132
  955. transformers/models/owlvit/processing_owlvit.py +48 -20
  956. transformers/models/paligemma/configuration_paligemma.py +1 -4
  957. transformers/models/paligemma/modeling_paligemma.py +93 -103
  958. transformers/models/paligemma/processing_paligemma.py +66 -13
  959. transformers/models/parakeet/configuration_parakeet.py +14 -7
  960. transformers/models/parakeet/feature_extraction_parakeet.py +12 -10
  961. transformers/models/parakeet/modeling_parakeet.py +28 -32
  962. transformers/models/parakeet/modular_parakeet.py +20 -23
  963. transformers/models/parakeet/processing_parakeet.py +5 -13
  964. transformers/models/parakeet/{tokenization_parakeet.py → tokenization_parakeet_fast.py} +7 -5
  965. transformers/models/patchtsmixer/configuration_patchtsmixer.py +8 -5
  966. transformers/models/patchtsmixer/modeling_patchtsmixer.py +62 -70
  967. transformers/models/patchtst/configuration_patchtst.py +9 -6
  968. transformers/models/patchtst/modeling_patchtst.py +80 -97
  969. transformers/models/pegasus/configuration_pegasus.py +5 -8
  970. transformers/models/pegasus/modeling_pegasus.py +66 -72
  971. transformers/models/pegasus/tokenization_pegasus.py +45 -15
  972. transformers/models/pegasus_x/configuration_pegasus_x.py +4 -5
  973. transformers/models/pegasus_x/modeling_pegasus_x.py +52 -55
  974. transformers/models/perceiver/configuration_perceiver.py +1 -0
  975. transformers/models/perceiver/image_processing_perceiver.py +25 -22
  976. transformers/models/perceiver/image_processing_perceiver_fast.py +9 -7
  977. transformers/models/perceiver/modeling_perceiver.py +146 -165
  978. transformers/models/perceiver/tokenization_perceiver.py +6 -3
  979. transformers/models/perception_lm/configuration_perception_lm.py +1 -0
  980. transformers/models/perception_lm/image_processing_perception_lm_fast.py +10 -8
  981. transformers/models/perception_lm/modeling_perception_lm.py +70 -71
  982. transformers/models/perception_lm/modular_perception_lm.py +61 -65
  983. transformers/models/perception_lm/processing_perception_lm.py +47 -13
  984. transformers/models/perception_lm/video_processing_perception_lm.py +1 -0
  985. transformers/models/persimmon/configuration_persimmon.py +28 -23
  986. transformers/models/persimmon/modeling_persimmon.py +45 -43
  987. transformers/models/phi/configuration_phi.py +28 -23
  988. transformers/models/phi/modeling_phi.py +43 -40
  989. transformers/models/phi/modular_phi.py +24 -23
  990. transformers/models/phi3/configuration_phi3.py +33 -28
  991. transformers/models/phi3/modeling_phi3.py +38 -36
  992. transformers/models/phi3/modular_phi3.py +17 -13
  993. transformers/models/phi4_multimodal/configuration_phi4_multimodal.py +33 -30
  994. transformers/models/phi4_multimodal/feature_extraction_phi4_multimodal.py +9 -7
  995. transformers/models/phi4_multimodal/image_processing_phi4_multimodal_fast.py +11 -11
  996. transformers/models/phi4_multimodal/modeling_phi4_multimodal.py +78 -95
  997. transformers/models/phi4_multimodal/modular_phi4_multimodal.py +80 -98
  998. transformers/models/phi4_multimodal/processing_phi4_multimodal.py +44 -7
  999. transformers/models/phimoe/configuration_phimoe.py +36 -31
  1000. transformers/models/phimoe/modeling_phimoe.py +45 -50
  1001. transformers/models/phimoe/modular_phimoe.py +4 -3
  1002. transformers/models/phobert/tokenization_phobert.py +6 -4
  1003. transformers/models/pix2struct/configuration_pix2struct.py +10 -12
  1004. transformers/models/pix2struct/image_processing_pix2struct.py +19 -15
  1005. transformers/models/pix2struct/image_processing_pix2struct_fast.py +15 -12
  1006. transformers/models/pix2struct/modeling_pix2struct.py +52 -58
  1007. transformers/models/pix2struct/processing_pix2struct.py +30 -5
  1008. transformers/models/pixtral/configuration_pixtral.py +14 -11
  1009. transformers/models/pixtral/image_processing_pixtral.py +28 -26
  1010. transformers/models/pixtral/image_processing_pixtral_fast.py +11 -10
  1011. transformers/models/pixtral/modeling_pixtral.py +34 -28
  1012. transformers/models/pixtral/processing_pixtral.py +53 -21
  1013. transformers/models/plbart/configuration_plbart.py +5 -8
  1014. transformers/models/plbart/modeling_plbart.py +106 -119
  1015. transformers/models/plbart/modular_plbart.py +33 -39
  1016. transformers/models/plbart/tokenization_plbart.py +7 -4
  1017. transformers/models/poolformer/configuration_poolformer.py +1 -0
  1018. transformers/models/poolformer/image_processing_poolformer.py +24 -21
  1019. transformers/models/poolformer/image_processing_poolformer_fast.py +15 -13
  1020. transformers/models/poolformer/modeling_poolformer.py +13 -23
  1021. transformers/models/pop2piano/configuration_pop2piano.py +8 -7
  1022. transformers/models/pop2piano/feature_extraction_pop2piano.py +9 -6
  1023. transformers/models/pop2piano/modeling_pop2piano.py +24 -26
  1024. transformers/models/pop2piano/processing_pop2piano.py +33 -25
  1025. transformers/models/pop2piano/tokenization_pop2piano.py +23 -15
  1026. transformers/models/prompt_depth_anything/configuration_prompt_depth_anything.py +3 -3
  1027. transformers/models/prompt_depth_anything/image_processing_prompt_depth_anything.py +28 -28
  1028. transformers/models/prompt_depth_anything/image_processing_prompt_depth_anything_fast.py +21 -20
  1029. transformers/models/prompt_depth_anything/modeling_prompt_depth_anything.py +13 -16
  1030. transformers/models/prompt_depth_anything/modular_prompt_depth_anything.py +13 -16
  1031. transformers/models/prophetnet/configuration_prophetnet.py +38 -37
  1032. transformers/models/prophetnet/modeling_prophetnet.py +131 -114
  1033. transformers/models/prophetnet/tokenization_prophetnet.py +16 -14
  1034. transformers/models/pvt/configuration_pvt.py +1 -0
  1035. transformers/models/pvt/image_processing_pvt.py +27 -24
  1036. transformers/models/pvt/image_processing_pvt_fast.py +2 -1
  1037. transformers/models/pvt/modeling_pvt.py +21 -21
  1038. transformers/models/pvt_v2/configuration_pvt_v2.py +4 -2
  1039. transformers/models/pvt_v2/modeling_pvt_v2.py +25 -28
  1040. transformers/models/qwen2/configuration_qwen2.py +25 -32
  1041. transformers/models/qwen2/modeling_qwen2.py +38 -36
  1042. transformers/models/qwen2/modular_qwen2.py +12 -11
  1043. transformers/models/qwen2/tokenization_qwen2.py +23 -12
  1044. transformers/models/qwen2_5_omni/configuration_qwen2_5_omni.py +26 -32
  1045. transformers/models/qwen2_5_omni/modeling_qwen2_5_omni.py +277 -340
  1046. transformers/models/qwen2_5_omni/modular_qwen2_5_omni.py +211 -278
  1047. transformers/models/qwen2_5_omni/processing_qwen2_5_omni.py +49 -41
  1048. transformers/models/qwen2_5_vl/configuration_qwen2_5_vl.py +35 -29
  1049. transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py +148 -203
  1050. transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py +118 -93
  1051. transformers/models/qwen2_5_vl/processing_qwen2_5_vl.py +43 -7
  1052. transformers/models/qwen2_audio/configuration_qwen2_audio.py +1 -0
  1053. transformers/models/qwen2_audio/modeling_qwen2_audio.py +40 -40
  1054. transformers/models/qwen2_audio/processing_qwen2_audio.py +42 -13
  1055. transformers/models/qwen2_moe/configuration_qwen2_moe.py +35 -42
  1056. transformers/models/qwen2_moe/modeling_qwen2_moe.py +46 -51
  1057. transformers/models/qwen2_moe/modular_qwen2_moe.py +10 -7
  1058. transformers/models/qwen2_vl/configuration_qwen2_vl.py +34 -29
  1059. transformers/models/qwen2_vl/image_processing_qwen2_vl.py +42 -41
  1060. transformers/models/qwen2_vl/image_processing_qwen2_vl_fast.py +15 -12
  1061. transformers/models/qwen2_vl/modeling_qwen2_vl.py +153 -199
  1062. transformers/models/qwen2_vl/processing_qwen2_vl.py +44 -7
  1063. transformers/models/qwen2_vl/video_processing_qwen2_vl.py +18 -38
  1064. transformers/models/qwen3/configuration_qwen3.py +27 -34
  1065. transformers/models/qwen3/modeling_qwen3.py +39 -36
  1066. transformers/models/qwen3/modular_qwen3.py +6 -4
  1067. transformers/models/qwen3_moe/configuration_qwen3_moe.py +32 -39
  1068. transformers/models/qwen3_moe/modeling_qwen3_moe.py +46 -51
  1069. transformers/models/qwen3_moe/modular_qwen3_moe.py +13 -10
  1070. transformers/models/qwen3_next/configuration_qwen3_next.py +35 -45
  1071. transformers/models/qwen3_next/modeling_qwen3_next.py +51 -47
  1072. transformers/models/qwen3_next/modular_qwen3_next.py +35 -34
  1073. transformers/models/qwen3_omni_moe/configuration_qwen3_omni_moe.py +101 -135
  1074. transformers/models/qwen3_omni_moe/modeling_qwen3_omni_moe.py +252 -355
  1075. transformers/models/qwen3_omni_moe/modular_qwen3_omni_moe.py +196 -250
  1076. transformers/models/qwen3_omni_moe/processing_qwen3_omni_moe.py +48 -40
  1077. transformers/models/qwen3_vl/configuration_qwen3_vl.py +29 -27
  1078. transformers/models/qwen3_vl/modeling_qwen3_vl.py +155 -233
  1079. transformers/models/qwen3_vl/modular_qwen3_vl.py +179 -206
  1080. transformers/models/qwen3_vl/processing_qwen3_vl.py +42 -6
  1081. transformers/models/qwen3_vl/video_processing_qwen3_vl.py +12 -10
  1082. transformers/models/qwen3_vl_moe/configuration_qwen3_vl_moe.py +30 -23
  1083. transformers/models/qwen3_vl_moe/modeling_qwen3_vl_moe.py +303 -358
  1084. transformers/models/qwen3_vl_moe/modular_qwen3_vl_moe.py +124 -87
  1085. transformers/models/rag/configuration_rag.py +15 -6
  1086. transformers/models/rag/modeling_rag.py +130 -127
  1087. transformers/models/rag/retrieval_rag.py +5 -3
  1088. transformers/models/rag/tokenization_rag.py +50 -0
  1089. transformers/models/recurrent_gemma/configuration_recurrent_gemma.py +30 -29
  1090. transformers/models/recurrent_gemma/modeling_recurrent_gemma.py +42 -53
  1091. transformers/models/reformer/configuration_reformer.py +8 -7
  1092. transformers/models/reformer/modeling_reformer.py +69 -80
  1093. transformers/models/reformer/tokenization_reformer.py +31 -11
  1094. transformers/models/regnet/configuration_regnet.py +1 -0
  1095. transformers/models/regnet/modeling_regnet.py +8 -15
  1096. transformers/models/rembert/configuration_rembert.py +2 -8
  1097. transformers/models/rembert/modeling_rembert.py +111 -121
  1098. transformers/models/rembert/tokenization_rembert.py +12 -2
  1099. transformers/models/resnet/configuration_resnet.py +1 -0
  1100. transformers/models/resnet/modeling_resnet.py +13 -27
  1101. transformers/models/roberta/configuration_roberta.py +3 -11
  1102. transformers/models/roberta/modeling_roberta.py +93 -94
  1103. transformers/models/roberta/modular_roberta.py +58 -58
  1104. transformers/models/roberta/tokenization_roberta.py +29 -17
  1105. transformers/models/roberta/tokenization_roberta_old.py +4 -2
  1106. transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py +3 -11
  1107. transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py +93 -94
  1108. transformers/models/roc_bert/configuration_roc_bert.py +2 -8
  1109. transformers/models/roc_bert/modeling_roc_bert.py +121 -122
  1110. transformers/models/roc_bert/tokenization_roc_bert.py +94 -88
  1111. transformers/models/roformer/configuration_roformer.py +3 -13
  1112. transformers/models/roformer/modeling_roformer.py +81 -85
  1113. transformers/models/roformer/tokenization_roformer.py +412 -74
  1114. transformers/models/roformer/tokenization_roformer_fast.py +160 -0
  1115. transformers/models/roformer/tokenization_utils.py +1 -0
  1116. transformers/models/rt_detr/configuration_rt_detr.py +2 -1
  1117. transformers/models/rt_detr/configuration_rt_detr_resnet.py +1 -0
  1118. transformers/models/rt_detr/image_processing_rt_detr.py +55 -54
  1119. transformers/models/rt_detr/image_processing_rt_detr_fast.py +26 -26
  1120. transformers/models/rt_detr/modeling_rt_detr.py +90 -99
  1121. transformers/models/rt_detr/modeling_rt_detr_resnet.py +6 -13
  1122. transformers/models/rt_detr/modular_rt_detr.py +16 -16
  1123. transformers/models/rt_detr_v2/configuration_rt_detr_v2.py +4 -6
  1124. transformers/models/rt_detr_v2/modeling_rt_detr_v2.py +90 -101
  1125. transformers/models/rt_detr_v2/modular_rt_detr_v2.py +12 -19
  1126. transformers/models/rwkv/configuration_rwkv.py +4 -2
  1127. transformers/models/rwkv/modeling_rwkv.py +32 -31
  1128. transformers/models/sam/configuration_sam.py +1 -3
  1129. transformers/models/sam/image_processing_sam.py +60 -59
  1130. transformers/models/sam/image_processing_sam_fast.py +27 -25
  1131. transformers/models/sam/modeling_sam.py +41 -47
  1132. transformers/models/sam/processing_sam.py +27 -39
  1133. transformers/models/sam2/configuration_sam2.py +3 -2
  1134. transformers/models/sam2/image_processing_sam2_fast.py +15 -14
  1135. transformers/models/sam2/modeling_sam2.py +90 -96
  1136. transformers/models/sam2/modular_sam2.py +91 -86
  1137. transformers/models/sam2/processing_sam2.py +47 -31
  1138. transformers/models/sam2_video/configuration_sam2_video.py +1 -0
  1139. transformers/models/sam2_video/modeling_sam2_video.py +144 -151
  1140. transformers/models/sam2_video/modular_sam2_video.py +104 -101
  1141. transformers/models/sam2_video/processing_sam2_video.py +66 -49
  1142. transformers/models/sam2_video/video_processing_sam2_video.py +4 -1
  1143. transformers/models/sam3/configuration_sam3.py +2 -21
  1144. transformers/models/sam3/image_processing_sam3_fast.py +20 -17
  1145. transformers/models/sam3/modeling_sam3.py +170 -184
  1146. transformers/models/sam3/modular_sam3.py +8 -3
  1147. transformers/models/sam3/processing_sam3.py +52 -37
  1148. transformers/models/sam3_tracker/__init__.py +1 -0
  1149. transformers/models/sam3_tracker/configuration_sam3_tracker.py +3 -1
  1150. transformers/models/sam3_tracker/modeling_sam3_tracker.py +77 -82
  1151. transformers/models/sam3_tracker/modular_sam3_tracker.py +3 -8
  1152. transformers/models/sam3_tracker/processing_sam3_tracker.py +48 -31
  1153. transformers/models/sam3_tracker_video/__init__.py +1 -0
  1154. transformers/models/sam3_tracker_video/configuration_sam3_tracker_video.py +1 -25
  1155. transformers/models/sam3_tracker_video/modeling_sam3_tracker_video.py +122 -135
  1156. transformers/models/sam3_tracker_video/modular_sam3_tracker_video.py +26 -35
  1157. transformers/models/sam3_tracker_video/processing_sam3_tracker_video.py +66 -50
  1158. transformers/models/sam3_video/configuration_sam3_video.py +1 -14
  1159. transformers/models/sam3_video/modeling_sam3_video.py +34 -33
  1160. transformers/models/sam3_video/processing_sam3_video.py +46 -26
  1161. transformers/models/sam_hq/__init__.py +1 -1
  1162. transformers/models/sam_hq/configuration_sam_hq.py +1 -3
  1163. transformers/models/sam_hq/modeling_sam_hq.py +69 -74
  1164. transformers/models/sam_hq/modular_sam_hq.py +25 -23
  1165. transformers/models/sam_hq/{processing_sam_hq.py → processing_samhq.py} +29 -41
  1166. transformers/models/seamless_m4t/configuration_seamless_m4t.py +10 -8
  1167. transformers/models/seamless_m4t/feature_extraction_seamless_m4t.py +11 -8
  1168. transformers/models/seamless_m4t/modeling_seamless_m4t.py +194 -212
  1169. transformers/models/seamless_m4t/processing_seamless_m4t.py +39 -18
  1170. transformers/models/seamless_m4t/tokenization_seamless_m4t.py +77 -40
  1171. transformers/models/seamless_m4t_v2/configuration_seamless_m4t_v2.py +10 -8
  1172. transformers/models/seamless_m4t_v2/modeling_seamless_m4t_v2.py +196 -204
  1173. transformers/models/seed_oss/configuration_seed_oss.py +32 -28
  1174. transformers/models/seed_oss/modeling_seed_oss.py +35 -33
  1175. transformers/models/seed_oss/modular_seed_oss.py +4 -3
  1176. transformers/models/segformer/configuration_segformer.py +10 -0
  1177. transformers/models/segformer/image_processing_segformer.py +42 -39
  1178. transformers/models/segformer/image_processing_segformer_fast.py +12 -10
  1179. transformers/models/segformer/modeling_segformer.py +31 -34
  1180. transformers/models/segformer/modular_segformer.py +10 -8
  1181. transformers/models/seggpt/configuration_seggpt.py +1 -0
  1182. transformers/models/seggpt/image_processing_seggpt.py +41 -38
  1183. transformers/models/seggpt/modeling_seggpt.py +38 -50
  1184. transformers/models/sew/configuration_sew.py +2 -4
  1185. transformers/models/sew/modeling_sew.py +36 -38
  1186. transformers/models/sew/modular_sew.py +13 -13
  1187. transformers/models/sew_d/configuration_sew_d.py +2 -4
  1188. transformers/models/sew_d/modeling_sew_d.py +30 -31
  1189. transformers/models/shieldgemma2/configuration_shieldgemma2.py +1 -0
  1190. transformers/models/shieldgemma2/modeling_shieldgemma2.py +17 -16
  1191. transformers/models/shieldgemma2/processing_shieldgemma2.py +5 -3
  1192. transformers/models/siglip/configuration_siglip.py +2 -4
  1193. transformers/models/siglip/image_processing_siglip.py +20 -17
  1194. transformers/models/siglip/image_processing_siglip_fast.py +1 -0
  1195. transformers/models/siglip/modeling_siglip.py +75 -84
  1196. transformers/models/siglip/processing_siglip.py +14 -2
  1197. transformers/models/siglip/tokenization_siglip.py +7 -6
  1198. transformers/models/siglip2/configuration_siglip2.py +2 -5
  1199. transformers/models/siglip2/image_processing_siglip2.py +16 -15
  1200. transformers/models/siglip2/image_processing_siglip2_fast.py +7 -6
  1201. transformers/models/siglip2/modeling_siglip2.py +129 -143
  1202. transformers/models/siglip2/modular_siglip2.py +46 -47
  1203. transformers/models/siglip2/processing_siglip2.py +14 -2
  1204. transformers/models/smollm3/configuration_smollm3.py +32 -29
  1205. transformers/models/smollm3/modeling_smollm3.py +39 -36
  1206. transformers/models/smollm3/modular_smollm3.py +35 -33
  1207. transformers/models/smolvlm/configuration_smolvlm.py +4 -2
  1208. transformers/models/smolvlm/image_processing_smolvlm.py +43 -42
  1209. transformers/models/smolvlm/image_processing_smolvlm_fast.py +15 -41
  1210. transformers/models/smolvlm/modeling_smolvlm.py +94 -126
  1211. transformers/models/smolvlm/modular_smolvlm.py +39 -50
  1212. transformers/models/smolvlm/processing_smolvlm.py +83 -15
  1213. transformers/models/smolvlm/video_processing_smolvlm.py +18 -16
  1214. transformers/models/speech_encoder_decoder/configuration_speech_encoder_decoder.py +1 -0
  1215. transformers/models/speech_encoder_decoder/modeling_speech_encoder_decoder.py +27 -26
  1216. transformers/models/speech_to_text/configuration_speech_to_text.py +9 -9
  1217. transformers/models/speech_to_text/feature_extraction_speech_to_text.py +13 -10
  1218. transformers/models/speech_to_text/modeling_speech_to_text.py +54 -66
  1219. transformers/models/speech_to_text/processing_speech_to_text.py +30 -4
  1220. transformers/models/speech_to_text/tokenization_speech_to_text.py +6 -5
  1221. transformers/models/speecht5/configuration_speecht5.py +9 -7
  1222. transformers/models/speecht5/feature_extraction_speecht5.py +37 -16
  1223. transformers/models/speecht5/modeling_speecht5.py +175 -213
  1224. transformers/models/speecht5/number_normalizer.py +1 -0
  1225. transformers/models/speecht5/processing_speecht5.py +37 -3
  1226. transformers/models/speecht5/tokenization_speecht5.py +5 -4
  1227. transformers/models/splinter/configuration_splinter.py +7 -6
  1228. transformers/models/splinter/modeling_splinter.py +59 -71
  1229. transformers/models/splinter/tokenization_splinter.py +30 -9
  1230. transformers/models/squeezebert/configuration_squeezebert.py +2 -14
  1231. transformers/models/squeezebert/modeling_squeezebert.py +62 -68
  1232. transformers/models/squeezebert/tokenization_squeezebert.py +1 -0
  1233. transformers/models/stablelm/configuration_stablelm.py +29 -24
  1234. transformers/models/stablelm/modeling_stablelm.py +45 -44
  1235. transformers/models/starcoder2/configuration_starcoder2.py +27 -30
  1236. transformers/models/starcoder2/modeling_starcoder2.py +41 -39
  1237. transformers/models/starcoder2/modular_starcoder2.py +16 -14
  1238. transformers/models/superglue/configuration_superglue.py +3 -7
  1239. transformers/models/superglue/image_processing_superglue.py +15 -15
  1240. transformers/models/superglue/image_processing_superglue_fast.py +10 -9
  1241. transformers/models/superglue/modeling_superglue.py +37 -42
  1242. transformers/models/superpoint/image_processing_superpoint.py +15 -15
  1243. transformers/models/superpoint/image_processing_superpoint_fast.py +11 -8
  1244. transformers/models/superpoint/modeling_superpoint.py +16 -18
  1245. transformers/models/swiftformer/configuration_swiftformer.py +1 -0
  1246. transformers/models/swiftformer/modeling_swiftformer.py +14 -18
  1247. transformers/models/swin/configuration_swin.py +1 -0
  1248. transformers/models/swin/modeling_swin.py +86 -86
  1249. transformers/models/swin2sr/configuration_swin2sr.py +1 -0
  1250. transformers/models/swin2sr/image_processing_swin2sr.py +13 -10
  1251. transformers/models/swin2sr/image_processing_swin2sr_fast.py +8 -4
  1252. transformers/models/swin2sr/modeling_swin2sr.py +63 -81
  1253. transformers/models/swinv2/configuration_swinv2.py +1 -0
  1254. transformers/models/swinv2/modeling_swinv2.py +104 -108
  1255. transformers/models/switch_transformers/configuration_switch_transformers.py +7 -11
  1256. transformers/models/switch_transformers/modeling_switch_transformers.py +44 -37
  1257. transformers/models/switch_transformers/modular_switch_transformers.py +41 -34
  1258. transformers/models/t5/configuration_t5.py +8 -14
  1259. transformers/models/t5/modeling_t5.py +92 -88
  1260. transformers/models/t5/tokenization_t5.py +9 -3
  1261. transformers/models/t5gemma/configuration_t5gemma.py +41 -43
  1262. transformers/models/t5gemma/modeling_t5gemma.py +107 -104
  1263. transformers/models/t5gemma/modular_t5gemma.py +120 -124
  1264. transformers/models/t5gemma2/configuration_t5gemma2.py +120 -80
  1265. transformers/models/t5gemma2/modeling_t5gemma2.py +125 -141
  1266. transformers/models/t5gemma2/modular_t5gemma2.py +104 -393
  1267. transformers/models/table_transformer/configuration_table_transformer.py +2 -1
  1268. transformers/models/table_transformer/modeling_table_transformer.py +49 -51
  1269. transformers/models/tapas/configuration_tapas.py +2 -12
  1270. transformers/models/tapas/modeling_tapas.py +67 -68
  1271. transformers/models/tapas/tokenization_tapas.py +153 -115
  1272. transformers/models/textnet/configuration_textnet.py +1 -0
  1273. transformers/models/textnet/image_processing_textnet.py +25 -22
  1274. transformers/models/textnet/image_processing_textnet_fast.py +10 -8
  1275. transformers/models/textnet/modeling_textnet.py +16 -28
  1276. transformers/models/time_series_transformer/configuration_time_series_transformer.py +8 -5
  1277. transformers/models/time_series_transformer/modeling_time_series_transformer.py +81 -83
  1278. transformers/models/timesfm/configuration_timesfm.py +1 -0
  1279. transformers/models/timesfm/modeling_timesfm.py +22 -33
  1280. transformers/models/timesfm/modular_timesfm.py +21 -32
  1281. transformers/models/timesformer/configuration_timesformer.py +1 -0
  1282. transformers/models/timesformer/modeling_timesformer.py +16 -15
  1283. transformers/models/timm_backbone/configuration_timm_backbone.py +1 -0
  1284. transformers/models/timm_backbone/modeling_timm_backbone.py +15 -17
  1285. transformers/models/timm_wrapper/configuration_timm_wrapper.py +3 -5
  1286. transformers/models/timm_wrapper/image_processing_timm_wrapper.py +5 -4
  1287. transformers/models/timm_wrapper/modeling_timm_wrapper.py +29 -34
  1288. transformers/models/trocr/configuration_trocr.py +8 -11
  1289. transformers/models/trocr/modeling_trocr.py +44 -45
  1290. transformers/models/trocr/processing_trocr.py +25 -5
  1291. transformers/models/tvp/configuration_tvp.py +2 -5
  1292. transformers/models/tvp/image_processing_tvp.py +52 -50
  1293. transformers/models/tvp/image_processing_tvp_fast.py +15 -15
  1294. transformers/models/tvp/modeling_tvp.py +27 -27
  1295. transformers/models/tvp/processing_tvp.py +14 -2
  1296. transformers/models/udop/configuration_udop.py +7 -16
  1297. transformers/models/udop/modeling_udop.py +73 -71
  1298. transformers/models/udop/processing_udop.py +26 -7
  1299. transformers/models/udop/tokenization_udop.py +105 -84
  1300. transformers/models/umt5/configuration_umt5.py +7 -8
  1301. transformers/models/umt5/modeling_umt5.py +90 -94
  1302. transformers/models/unispeech/configuration_unispeech.py +2 -4
  1303. transformers/models/unispeech/modeling_unispeech.py +49 -51
  1304. transformers/models/unispeech/modular_unispeech.py +22 -22
  1305. transformers/models/unispeech_sat/configuration_unispeech_sat.py +2 -4
  1306. transformers/models/unispeech_sat/modeling_unispeech_sat.py +65 -69
  1307. transformers/models/unispeech_sat/modular_unispeech_sat.py +23 -23
  1308. transformers/models/univnet/feature_extraction_univnet.py +14 -14
  1309. transformers/models/univnet/modeling_univnet.py +8 -8
  1310. transformers/models/upernet/configuration_upernet.py +1 -0
  1311. transformers/models/upernet/modeling_upernet.py +13 -11
  1312. transformers/models/vaultgemma/__init__.py +1 -0
  1313. transformers/models/vaultgemma/configuration_vaultgemma.py +33 -29
  1314. transformers/models/vaultgemma/modeling_vaultgemma.py +41 -39
  1315. transformers/models/vaultgemma/modular_vaultgemma.py +31 -29
  1316. transformers/models/video_llama_3/configuration_video_llama_3.py +0 -4
  1317. transformers/models/video_llama_3/image_processing_video_llama_3.py +42 -43
  1318. transformers/models/video_llama_3/image_processing_video_llama_3_fast.py +14 -12
  1319. transformers/models/video_llama_3/modeling_video_llama_3.py +109 -157
  1320. transformers/models/video_llama_3/modular_video_llama_3.py +146 -155
  1321. transformers/models/video_llama_3/processing_video_llama_3.py +39 -5
  1322. transformers/models/video_llama_3/video_processing_video_llama_3.py +23 -42
  1323. transformers/models/video_llava/configuration_video_llava.py +1 -4
  1324. transformers/models/video_llava/image_processing_video_llava.py +38 -35
  1325. transformers/models/video_llava/modeling_video_llava.py +146 -146
  1326. transformers/models/video_llava/processing_video_llava.py +78 -38
  1327. transformers/models/video_llava/video_processing_video_llava.py +1 -0
  1328. transformers/models/videomae/configuration_videomae.py +1 -0
  1329. transformers/models/videomae/image_processing_videomae.py +34 -31
  1330. transformers/models/videomae/modeling_videomae.py +17 -14
  1331. transformers/models/videomae/video_processing_videomae.py +1 -0
  1332. transformers/models/vilt/configuration_vilt.py +4 -6
  1333. transformers/models/vilt/image_processing_vilt.py +30 -29
  1334. transformers/models/vilt/image_processing_vilt_fast.py +16 -15
  1335. transformers/models/vilt/modeling_vilt.py +90 -116
  1336. transformers/models/vilt/processing_vilt.py +14 -2
  1337. transformers/models/vipllava/configuration_vipllava.py +1 -4
  1338. transformers/models/vipllava/modeling_vipllava.py +70 -99
  1339. transformers/models/vipllava/modular_vipllava.py +54 -78
  1340. transformers/models/vision_encoder_decoder/configuration_vision_encoder_decoder.py +1 -0
  1341. transformers/models/vision_encoder_decoder/modeling_vision_encoder_decoder.py +27 -28
  1342. transformers/models/vision_text_dual_encoder/configuration_vision_text_dual_encoder.py +1 -0
  1343. transformers/models/vision_text_dual_encoder/modeling_vision_text_dual_encoder.py +41 -46
  1344. transformers/models/vision_text_dual_encoder/processing_vision_text_dual_encoder.py +16 -2
  1345. transformers/models/visual_bert/configuration_visual_bert.py +2 -6
  1346. transformers/models/visual_bert/modeling_visual_bert.py +92 -98
  1347. transformers/models/vit/configuration_vit.py +1 -0
  1348. transformers/models/vit/image_processing_vit.py +22 -19
  1349. transformers/models/vit/image_processing_vit_fast.py +1 -0
  1350. transformers/models/vit/modeling_vit.py +17 -17
  1351. transformers/models/vit_mae/configuration_vit_mae.py +1 -0
  1352. transformers/models/vit_mae/modeling_vit_mae.py +27 -29
  1353. transformers/models/vit_msn/configuration_vit_msn.py +1 -0
  1354. transformers/models/vit_msn/modeling_vit_msn.py +16 -18
  1355. transformers/models/vitdet/configuration_vitdet.py +1 -0
  1356. transformers/models/vitdet/modeling_vitdet.py +14 -14
  1357. transformers/models/vitmatte/configuration_vitmatte.py +5 -2
  1358. transformers/models/vitmatte/image_processing_vitmatte.py +18 -15
  1359. transformers/models/vitmatte/image_processing_vitmatte_fast.py +18 -16
  1360. transformers/models/vitmatte/modeling_vitmatte.py +11 -14
  1361. transformers/models/vitpose/configuration_vitpose.py +7 -4
  1362. transformers/models/vitpose/image_processing_vitpose.py +25 -24
  1363. transformers/models/vitpose/image_processing_vitpose_fast.py +11 -9
  1364. transformers/models/vitpose/modeling_vitpose.py +14 -14
  1365. transformers/models/vitpose_backbone/configuration_vitpose_backbone.py +1 -0
  1366. transformers/models/vitpose_backbone/modeling_vitpose_backbone.py +10 -8
  1367. transformers/models/vits/configuration_vits.py +1 -4
  1368. transformers/models/vits/modeling_vits.py +42 -44
  1369. transformers/models/vits/tokenization_vits.py +4 -3
  1370. transformers/models/vivit/configuration_vivit.py +1 -0
  1371. transformers/models/vivit/image_processing_vivit.py +39 -36
  1372. transformers/models/vivit/modeling_vivit.py +8 -6
  1373. transformers/models/vjepa2/__init__.py +1 -0
  1374. transformers/models/vjepa2/configuration_vjepa2.py +1 -0
  1375. transformers/models/vjepa2/modeling_vjepa2.py +32 -31
  1376. transformers/models/vjepa2/video_processing_vjepa2.py +1 -0
  1377. transformers/models/voxtral/__init__.py +1 -0
  1378. transformers/models/voxtral/configuration_voxtral.py +2 -0
  1379. transformers/models/voxtral/modeling_voxtral.py +47 -40
  1380. transformers/models/voxtral/modular_voxtral.py +40 -37
  1381. transformers/models/voxtral/processing_voxtral.py +48 -25
  1382. transformers/models/wav2vec2/configuration_wav2vec2.py +2 -4
  1383. transformers/models/wav2vec2/feature_extraction_wav2vec2.py +10 -7
  1384. transformers/models/wav2vec2/modeling_wav2vec2.py +121 -73
  1385. transformers/models/wav2vec2/processing_wav2vec2.py +35 -6
  1386. transformers/models/wav2vec2/tokenization_wav2vec2.py +332 -20
  1387. transformers/models/wav2vec2_bert/configuration_wav2vec2_bert.py +2 -4
  1388. transformers/models/wav2vec2_bert/modeling_wav2vec2_bert.py +62 -70
  1389. transformers/models/wav2vec2_bert/modular_wav2vec2_bert.py +48 -57
  1390. transformers/models/wav2vec2_bert/processing_wav2vec2_bert.py +35 -6
  1391. transformers/models/wav2vec2_conformer/configuration_wav2vec2_conformer.py +2 -4
  1392. transformers/models/wav2vec2_conformer/modeling_wav2vec2_conformer.py +77 -90
  1393. transformers/models/wav2vec2_conformer/modular_wav2vec2_conformer.py +30 -37
  1394. transformers/models/wav2vec2_phoneme/tokenization_wav2vec2_phoneme.py +17 -16
  1395. transformers/models/wav2vec2_with_lm/processing_wav2vec2_with_lm.py +55 -36
  1396. transformers/models/wavlm/configuration_wavlm.py +2 -4
  1397. transformers/models/wavlm/modeling_wavlm.py +48 -50
  1398. transformers/models/wavlm/modular_wavlm.py +5 -4
  1399. transformers/models/whisper/configuration_whisper.py +5 -6
  1400. transformers/models/whisper/english_normalizer.py +4 -3
  1401. transformers/models/whisper/feature_extraction_whisper.py +24 -9
  1402. transformers/models/whisper/generation_whisper.py +48 -26
  1403. transformers/models/whisper/modeling_whisper.py +73 -79
  1404. transformers/models/whisper/processing_whisper.py +20 -3
  1405. transformers/models/whisper/tokenization_whisper.py +43 -11
  1406. transformers/models/x_clip/configuration_x_clip.py +2 -4
  1407. transformers/models/x_clip/modeling_x_clip.py +93 -96
  1408. transformers/models/x_clip/processing_x_clip.py +14 -2
  1409. transformers/models/xcodec/configuration_xcodec.py +6 -4
  1410. transformers/models/xcodec/modeling_xcodec.py +17 -20
  1411. transformers/models/xglm/configuration_xglm.py +8 -9
  1412. transformers/models/xglm/modeling_xglm.py +55 -60
  1413. transformers/models/xglm/tokenization_xglm.py +11 -3
  1414. transformers/models/xlm/configuration_xlm.py +8 -10
  1415. transformers/models/xlm/modeling_xlm.py +144 -144
  1416. transformers/models/xlm/tokenization_xlm.py +5 -3
  1417. transformers/models/xlm_roberta/configuration_xlm_roberta.py +3 -11
  1418. transformers/models/xlm_roberta/modeling_xlm_roberta.py +194 -195
  1419. transformers/models/xlm_roberta/modular_xlm_roberta.py +53 -50
  1420. transformers/models/xlm_roberta/tokenization_xlm_roberta.py +18 -8
  1421. transformers/models/xlm_roberta_xl/configuration_xlm_roberta_xl.py +2 -10
  1422. transformers/models/xlm_roberta_xl/modeling_xlm_roberta_xl.py +93 -94
  1423. transformers/models/xlm_roberta_xl/modular_xlm_roberta_xl.py +70 -67
  1424. transformers/models/xlnet/configuration_xlnet.py +12 -3
  1425. transformers/models/xlnet/modeling_xlnet.py +163 -152
  1426. transformers/models/xlnet/tokenization_xlnet.py +9 -2
  1427. transformers/models/xlstm/configuration_xlstm.py +12 -8
  1428. transformers/models/xlstm/modeling_xlstm.py +65 -62
  1429. transformers/models/xmod/configuration_xmod.py +3 -11
  1430. transformers/models/xmod/modeling_xmod.py +110 -108
  1431. transformers/models/yolos/configuration_yolos.py +1 -0
  1432. transformers/models/yolos/image_processing_yolos.py +62 -60
  1433. transformers/models/yolos/image_processing_yolos_fast.py +45 -42
  1434. transformers/models/yolos/modeling_yolos.py +16 -16
  1435. transformers/models/yolos/modular_yolos.py +19 -17
  1436. transformers/models/yoso/configuration_yoso.py +2 -8
  1437. transformers/models/yoso/modeling_yoso.py +63 -70
  1438. transformers/models/zamba/configuration_zamba.py +8 -5
  1439. transformers/models/zamba/modeling_zamba.py +78 -81
  1440. transformers/models/zamba2/configuration_zamba2.py +50 -44
  1441. transformers/models/zamba2/modeling_zamba2.py +97 -97
  1442. transformers/models/zamba2/modular_zamba2.py +48 -46
  1443. transformers/models/zoedepth/configuration_zoedepth.py +2 -1
  1444. transformers/models/zoedepth/image_processing_zoedepth.py +29 -28
  1445. transformers/models/zoedepth/image_processing_zoedepth_fast.py +24 -21
  1446. transformers/models/zoedepth/modeling_zoedepth.py +18 -26
  1447. transformers/pipelines/__init__.py +114 -57
  1448. transformers/pipelines/any_to_any.py +22 -14
  1449. transformers/pipelines/audio_utils.py +2 -1
  1450. transformers/pipelines/automatic_speech_recognition.py +12 -20
  1451. transformers/pipelines/base.py +27 -15
  1452. transformers/{models/pe_audio/processing_pe_audio.py → pipelines/deprecated/__init__.py} +3 -10
  1453. transformers/pipelines/deprecated/text2text_generation.py +408 -0
  1454. transformers/pipelines/document_question_answering.py +2 -4
  1455. transformers/pipelines/image_text_to_text.py +1 -0
  1456. transformers/pipelines/image_to_text.py +229 -0
  1457. transformers/pipelines/question_answering.py +44 -5
  1458. transformers/pipelines/text_classification.py +14 -1
  1459. transformers/pipelines/text_generation.py +1 -1
  1460. transformers/pipelines/text_to_audio.py +2 -2
  1461. transformers/pipelines/token_classification.py +22 -1
  1462. transformers/pipelines/video_classification.py +9 -1
  1463. transformers/pipelines/zero_shot_audio_classification.py +1 -0
  1464. transformers/pipelines/zero_shot_classification.py +6 -0
  1465. transformers/pipelines/zero_shot_image_classification.py +7 -0
  1466. transformers/processing_utils.py +145 -230
  1467. transformers/quantizers/auto.py +4 -2
  1468. transformers/quantizers/base.py +173 -53
  1469. transformers/quantizers/quantizer_aqlm.py +23 -2
  1470. transformers/quantizers/quantizer_auto_round.py +12 -2
  1471. transformers/quantizers/quantizer_awq.py +89 -20
  1472. transformers/quantizers/quantizer_bitnet.py +14 -4
  1473. transformers/quantizers/quantizer_bnb_4bit.py +155 -18
  1474. transformers/quantizers/quantizer_bnb_8bit.py +110 -24
  1475. transformers/quantizers/quantizer_compressed_tensors.py +9 -2
  1476. transformers/quantizers/quantizer_eetq.py +74 -16
  1477. transformers/quantizers/quantizer_fbgemm_fp8.py +138 -38
  1478. transformers/quantizers/quantizer_finegrained_fp8.py +113 -26
  1479. transformers/quantizers/quantizer_fp_quant.py +82 -52
  1480. transformers/quantizers/quantizer_gptq.py +28 -8
  1481. transformers/quantizers/quantizer_higgs.py +60 -42
  1482. transformers/quantizers/quantizer_hqq.py +153 -144
  1483. transformers/quantizers/quantizer_mxfp4.py +194 -14
  1484. transformers/quantizers/quantizer_quanto.py +79 -35
  1485. transformers/quantizers/quantizer_quark.py +18 -36
  1486. transformers/quantizers/quantizer_spqr.py +12 -4
  1487. transformers/quantizers/quantizer_torchao.py +325 -50
  1488. transformers/quantizers/quantizer_vptq.py +27 -4
  1489. transformers/quantizers/quantizers_utils.py +0 -20
  1490. transformers/safetensors_conversion.py +3 -9
  1491. transformers/testing_utils.py +82 -326
  1492. transformers/tokenization_mistral_common.py +903 -568
  1493. transformers/tokenization_utils_base.py +340 -220
  1494. transformers/tokenization_utils_sentencepiece.py +6 -5
  1495. transformers/tokenization_utils_tokenizers.py +113 -226
  1496. transformers/trainer.py +53 -60
  1497. transformers/trainer_callback.py +0 -8
  1498. transformers/trainer_seq2seq.py +1 -5
  1499. transformers/trainer_utils.py +1 -1
  1500. transformers/training_args.py +41 -77
  1501. transformers/utils/__init__.py +4 -8
  1502. transformers/utils/attention_visualizer.py +5 -5
  1503. transformers/utils/auto_docstring.py +37 -599
  1504. transformers/utils/doc.py +36 -4
  1505. transformers/utils/dummy_pt_objects.py +42 -0
  1506. transformers/utils/generic.py +28 -111
  1507. transformers/utils/hub.py +15 -5
  1508. transformers/utils/import_utils.py +32 -165
  1509. transformers/utils/kernel_config.py +19 -74
  1510. transformers/utils/loading_report.py +15 -25
  1511. transformers/utils/quantization_config.py +241 -72
  1512. transformers/video_processing_utils.py +39 -41
  1513. transformers/video_utils.py +22 -18
  1514. {transformers-5.0.0.dist-info → transformers-5.0.0rc0.dist-info}/METADATA +236 -284
  1515. transformers-5.0.0rc0.dist-info/RECORD +1987 -0
  1516. {transformers-5.0.0.dist-info → transformers-5.0.0rc0.dist-info}/WHEEL +1 -1
  1517. transformers/integrations/moe.py +0 -360
  1518. transformers/integrations/quark.py +0 -53
  1519. transformers/loss/loss_lw_detr.py +0 -356
  1520. transformers/models/ernie4_5_vl_moe/__init__.py +0 -31
  1521. transformers/models/ernie4_5_vl_moe/configuration_ernie4_5_vl_moe.py +0 -340
  1522. transformers/models/ernie4_5_vl_moe/image_processing_ernie4_5_vl_moe.py +0 -455
  1523. transformers/models/ernie4_5_vl_moe/image_processing_ernie4_5_vl_moe_fast.py +0 -231
  1524. transformers/models/ernie4_5_vl_moe/modeling_ernie4_5_vl_moe.py +0 -1936
  1525. transformers/models/ernie4_5_vl_moe/modular_ernie4_5_vl_moe.py +0 -1925
  1526. transformers/models/ernie4_5_vl_moe/processing_ernie4_5_vl_moe.py +0 -249
  1527. transformers/models/ernie4_5_vl_moe/video_processing_ernie4_5_vl_moe.py +0 -593
  1528. transformers/models/fast_vlm/__init__.py +0 -27
  1529. transformers/models/fast_vlm/configuration_fast_vlm.py +0 -137
  1530. transformers/models/fast_vlm/modeling_fast_vlm.py +0 -432
  1531. transformers/models/fast_vlm/modular_fast_vlm.py +0 -373
  1532. transformers/models/glm4_moe_lite/__init__.py +0 -28
  1533. transformers/models/glm4_moe_lite/configuration_glm4_moe_lite.py +0 -233
  1534. transformers/models/glm4_moe_lite/modeling_glm4_moe_lite.py +0 -740
  1535. transformers/models/glm4_moe_lite/modular_glm4_moe_lite.py +0 -302
  1536. transformers/models/glm_image/__init__.py +0 -31
  1537. transformers/models/glm_image/configuration_glm_image.py +0 -351
  1538. transformers/models/glm_image/image_processing_glm_image.py +0 -503
  1539. transformers/models/glm_image/image_processing_glm_image_fast.py +0 -294
  1540. transformers/models/glm_image/modeling_glm_image.py +0 -1642
  1541. transformers/models/glm_image/modular_glm_image.py +0 -1531
  1542. transformers/models/glm_image/processing_glm_image.py +0 -217
  1543. transformers/models/glmasr/__init__.py +0 -29
  1544. transformers/models/glmasr/configuration_glmasr.py +0 -196
  1545. transformers/models/glmasr/modeling_glmasr.py +0 -517
  1546. transformers/models/glmasr/modular_glmasr.py +0 -443
  1547. transformers/models/glmasr/processing_glmasr.py +0 -331
  1548. transformers/models/jais2/__init__.py +0 -27
  1549. transformers/models/jais2/configuration_jais2.py +0 -148
  1550. transformers/models/jais2/modeling_jais2.py +0 -484
  1551. transformers/models/jais2/modular_jais2.py +0 -194
  1552. transformers/models/lasr/__init__.py +0 -29
  1553. transformers/models/lasr/configuration_lasr.py +0 -244
  1554. transformers/models/lasr/feature_extraction_lasr.py +0 -275
  1555. transformers/models/lasr/modeling_lasr.py +0 -727
  1556. transformers/models/lasr/modular_lasr.py +0 -574
  1557. transformers/models/lasr/processing_lasr.py +0 -100
  1558. transformers/models/lasr/tokenization_lasr.py +0 -184
  1559. transformers/models/lighton_ocr/__init__.py +0 -28
  1560. transformers/models/lighton_ocr/configuration_lighton_ocr.py +0 -128
  1561. transformers/models/lighton_ocr/modeling_lighton_ocr.py +0 -463
  1562. transformers/models/lighton_ocr/modular_lighton_ocr.py +0 -404
  1563. transformers/models/lighton_ocr/processing_lighton_ocr.py +0 -229
  1564. transformers/models/lw_detr/__init__.py +0 -27
  1565. transformers/models/lw_detr/configuration_lw_detr.py +0 -374
  1566. transformers/models/lw_detr/modeling_lw_detr.py +0 -1702
  1567. transformers/models/lw_detr/modular_lw_detr.py +0 -1615
  1568. transformers/models/minimax_m2/__init__.py +0 -28
  1569. transformers/models/minimax_m2/configuration_minimax_m2.py +0 -188
  1570. transformers/models/minimax_m2/modeling_minimax_m2.py +0 -704
  1571. transformers/models/minimax_m2/modular_minimax_m2.py +0 -346
  1572. transformers/models/paddleocr_vl/__init__.py +0 -31
  1573. transformers/models/paddleocr_vl/configuration_paddleocr_vl.py +0 -335
  1574. transformers/models/paddleocr_vl/image_processing_paddleocr_vl.py +0 -503
  1575. transformers/models/paddleocr_vl/image_processing_paddleocr_vl_fast.py +0 -209
  1576. transformers/models/paddleocr_vl/modeling_paddleocr_vl.py +0 -1683
  1577. transformers/models/paddleocr_vl/modular_paddleocr_vl.py +0 -1380
  1578. transformers/models/paddleocr_vl/processing_paddleocr_vl.py +0 -133
  1579. transformers/models/pe_audio/__init__.py +0 -29
  1580. transformers/models/pe_audio/configuration_pe_audio.py +0 -204
  1581. transformers/models/pe_audio/feature_extraction_pe_audio.py +0 -160
  1582. transformers/models/pe_audio/modeling_pe_audio.py +0 -819
  1583. transformers/models/pe_audio/modular_pe_audio.py +0 -298
  1584. transformers/models/pe_audio_video/__init__.py +0 -28
  1585. transformers/models/pe_audio_video/configuration_pe_audio_video.py +0 -223
  1586. transformers/models/pe_audio_video/modeling_pe_audio_video.py +0 -971
  1587. transformers/models/pe_audio_video/modular_pe_audio_video.py +0 -763
  1588. transformers/models/pe_video/__init__.py +0 -29
  1589. transformers/models/pe_video/configuration_pe_video.py +0 -209
  1590. transformers/models/pe_video/modeling_pe_video.py +0 -647
  1591. transformers/models/pe_video/modular_pe_video.py +0 -231
  1592. transformers/models/pe_video/processing_pe_video.py +0 -10
  1593. transformers/models/pe_video/video_processing_pe_video.py +0 -64
  1594. transformers/models/pixio/__init__.py +0 -29
  1595. transformers/models/pixio/configuration_pixio.py +0 -150
  1596. transformers/models/pixio/modeling_pixio.py +0 -507
  1597. transformers/models/pixio/modular_pixio.py +0 -403
  1598. transformers/models/solar_open/__init__.py +0 -27
  1599. transformers/models/solar_open/configuration_solar_open.py +0 -184
  1600. transformers/models/solar_open/modeling_solar_open.py +0 -642
  1601. transformers/models/solar_open/modular_solar_open.py +0 -224
  1602. transformers/trainer_jit_checkpoint.py +0 -125
  1603. transformers-5.0.0.dist-info/RECORD +0 -2068
  1604. {transformers-5.0.0.dist-info/licenses → transformers-5.0.0rc0.dist-info}/LICENSE +0 -0
  1605. {transformers-5.0.0.dist-info → transformers-5.0.0rc0.dist-info}/entry_points.txt +0 -0
  1606. {transformers-5.0.0.dist-info → transformers-5.0.0rc0.dist-info}/top_level.txt +0 -0
transformers/modeling_outputs.py CHANGED
@@ -12,7 +12,9 @@
12
12
  # See the License for the specific language governing permissions and
13
13
  # limitations under the License.
14
14
 
15
+ import warnings
15
16
  from dataclasses import dataclass
17
+ from typing import Optional
16
18
 
17
19
  import torch
18
20
 
@@ -41,9 +43,9 @@ class BaseModelOutput(ModelOutput):
41
43
  heads.
42
44
  """
43
45
 
44
- last_hidden_state: torch.FloatTensor | None = None
45
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
46
- attentions: tuple[torch.FloatTensor, ...] | None = None
46
+ last_hidden_state: Optional[torch.FloatTensor] = None
47
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
48
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
47
49
 
48
50
 
49
51
  @dataclass
@@ -61,8 +63,8 @@ class BaseModelOutputWithNoAttention(ModelOutput):
61
63
  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
62
64
  """
63
65
 
64
- last_hidden_state: torch.FloatTensor | None = None
65
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
66
+ last_hidden_state: Optional[torch.FloatTensor] = None
67
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
66
68
 
67
69
 
68
70
  @dataclass
@@ -91,10 +93,10 @@ class BaseModelOutputWithPooling(ModelOutput):
91
93
  heads.
92
94
  """
93
95
 
94
- last_hidden_state: torch.FloatTensor | None = None
95
- pooler_output: torch.FloatTensor | None = None
96
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
97
- attentions: tuple[torch.FloatTensor, ...] | None = None
96
+ last_hidden_state: Optional[torch.FloatTensor] = None
97
+ pooler_output: Optional[torch.FloatTensor] = None
98
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
99
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
98
100
 
99
101
 
100
102
  @dataclass
@@ -114,9 +116,9 @@ class BaseModelOutputWithPoolingAndNoAttention(ModelOutput):
114
116
  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
115
117
  """
116
118
 
117
- last_hidden_state: torch.FloatTensor | None = None
118
- pooler_output: torch.FloatTensor | None = None
119
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
119
+ last_hidden_state: Optional[torch.FloatTensor] = None
120
+ pooler_output: Optional[torch.FloatTensor] = None
121
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
120
122
 
121
123
 
122
124
  @dataclass
@@ -149,10 +151,10 @@ class BaseModelOutputWithPast(ModelOutput):
149
151
  heads.
150
152
  """
151
153
 
152
- last_hidden_state: torch.FloatTensor | None = None
153
- past_key_values: Cache | None = None
154
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
155
- attentions: tuple[torch.FloatTensor, ...] | None = None
154
+ last_hidden_state: Optional[torch.FloatTensor] = None
155
+ past_key_values: Optional[Cache] = None
156
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
157
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
156
158
 
157
159
 
158
160
  @dataclass
@@ -182,10 +184,10 @@ class BaseModelOutputWithCrossAttentions(ModelOutput):
182
184
  weighted average in the cross-attention heads.
183
185
  """
184
186
 
185
- last_hidden_state: torch.FloatTensor | None = None
186
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
187
- attentions: tuple[torch.FloatTensor, ...] | None = None
188
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
187
+ last_hidden_state: Optional[torch.FloatTensor] = None
188
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
189
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
190
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
189
191
 
190
192
 
191
193
  @dataclass
@@ -226,12 +228,12 @@ class BaseModelOutputWithPoolingAndCrossAttentions(ModelOutput):
226
228
  input) to speed up sequential decoding.
227
229
  """
228
230
 
229
- last_hidden_state: torch.FloatTensor | None = None
230
- pooler_output: torch.FloatTensor | None = None
231
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
232
- past_key_values: Cache | None = None
233
- attentions: tuple[torch.FloatTensor, ...] | None = None
234
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
231
+ last_hidden_state: Optional[torch.FloatTensor] = None
232
+ pooler_output: Optional[torch.FloatTensor] = None
233
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
234
+ past_key_values: Optional[Cache] = None
235
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
236
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
235
237
 
236
238
 
237
239
  @dataclass
@@ -270,11 +272,11 @@ class BaseModelOutputWithPastAndCrossAttentions(ModelOutput):
270
272
  weighted average in the cross-attention heads.
271
273
  """
272
274
 
273
- last_hidden_state: torch.FloatTensor | None = None
274
- past_key_values: Cache | None = None
275
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
276
- attentions: tuple[torch.FloatTensor, ...] | None = None
277
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
275
+ last_hidden_state: Optional[torch.FloatTensor] = None
276
+ past_key_values: Optional[Cache] = None
277
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
278
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
279
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
278
280
 
279
281
 
280
282
  @dataclass
@@ -315,14 +317,14 @@ class MoECausalLMOutputWithPast(ModelOutput):
315
317
  modules.
316
318
  """
317
319
 
318
- loss: torch.FloatTensor | None = None
319
- logits: torch.FloatTensor | None = None
320
- past_key_values: Cache | None = None
321
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
322
- attentions: tuple[torch.FloatTensor, ...] | None = None
323
- z_loss: torch.FloatTensor | None = None
324
- aux_loss: torch.FloatTensor | None = None
325
- router_logits: tuple[torch.FloatTensor] | None = None
320
+ loss: Optional[torch.FloatTensor] = None
321
+ logits: Optional[torch.FloatTensor] = None
322
+ past_key_values: Optional[Cache] = None
323
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
324
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
325
+ z_loss: Optional[torch.FloatTensor] = None
326
+ aux_loss: Optional[torch.FloatTensor] = None
327
+ router_logits: Optional[tuple[torch.FloatTensor]] = None
326
328
 
327
329
 
328
330
  @dataclass
@@ -351,11 +353,11 @@ class MoEModelOutput(ModelOutput):
351
353
  loss and the z_loss for Mixture of Experts models.
352
354
  """
353
355
 
354
- last_hidden_state: torch.FloatTensor | None = None
355
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
356
- attentions: tuple[torch.FloatTensor, ...] | None = None
357
- router_probs: tuple[torch.FloatTensor] | None = None
358
- router_logits: tuple[torch.FloatTensor] | None = None
356
+ last_hidden_state: Optional[torch.FloatTensor] = None
357
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
358
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
359
+ router_probs: Optional[tuple[torch.FloatTensor]] = None
360
+ router_logits: Optional[tuple[torch.FloatTensor]] = None
359
361
 
360
362
 
361
363
  @dataclass
@@ -390,11 +392,11 @@ class MoeModelOutputWithPast(ModelOutput):
390
392
  loss for Mixture of Experts models.
391
393
  """
392
394
 
393
- last_hidden_state: torch.FloatTensor | None = None
394
- past_key_values: Cache | None = None
395
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
396
- attentions: tuple[torch.FloatTensor, ...] | None = None
397
- router_logits: tuple[torch.FloatTensor] | None = None
395
+ last_hidden_state: Optional[torch.FloatTensor] = None
396
+ past_key_values: Optional[Cache] = None
397
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
398
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
399
+ router_logits: Optional[tuple[torch.FloatTensor]] = None
398
400
 
399
401
 
400
402
  @dataclass
@@ -436,13 +438,13 @@ class MoeCausalLMOutputWithPast(ModelOutput):
436
438
  heads.
437
439
  """
438
440
 
439
- loss: torch.FloatTensor | None = None
440
- aux_loss: torch.FloatTensor | None = None
441
- logits: torch.FloatTensor | None = None
442
- past_key_values: Cache | None = None
443
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
444
- attentions: tuple[torch.FloatTensor, ...] | None = None
445
- router_logits: tuple[torch.FloatTensor] | None = None
441
+ loss: Optional[torch.FloatTensor] = None
442
+ aux_loss: Optional[torch.FloatTensor] = None
443
+ logits: Optional[torch.FloatTensor] = None
444
+ past_key_values: Optional[Cache] = None
445
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
446
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
447
+ router_logits: Optional[tuple[torch.FloatTensor]] = None
446
448
 
447
449
 
448
450
  @dataclass
@@ -487,13 +489,13 @@ class MoEModelOutputWithPastAndCrossAttentions(ModelOutput):
487
489
  loss and the z_loss for Mixture of Experts models.
488
490
  """
489
491
 
490
- last_hidden_state: torch.FloatTensor | None = None
491
- past_key_values: Cache | None = None
492
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
493
- attentions: tuple[torch.FloatTensor, ...] | None = None
494
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
495
- router_probs: tuple[torch.FloatTensor] | None = None
496
- router_logits: tuple[torch.FloatTensor] | None = None
492
+ last_hidden_state: Optional[torch.FloatTensor] = None
493
+ past_key_values: Optional[Cache] = None
494
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
495
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
496
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
497
+ router_probs: Optional[tuple[torch.FloatTensor]] = None
498
+ router_logits: Optional[tuple[torch.FloatTensor]] = None
497
499
 
498
500
 
499
501
  @dataclass
@@ -545,14 +547,14 @@ class Seq2SeqModelOutput(ModelOutput):
545
547
  self-attention heads.
546
548
  """
547
549
 
548
- last_hidden_state: torch.FloatTensor | None = None
549
- past_key_values: EncoderDecoderCache | None = None
550
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
551
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
552
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
553
- encoder_last_hidden_state: torch.FloatTensor | None = None
554
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
555
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
550
+ last_hidden_state: Optional[torch.FloatTensor] = None
551
+ past_key_values: Optional[EncoderDecoderCache] = None
552
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
553
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
554
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
555
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
556
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
557
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
556
558
 
557
559
 
558
560
  @dataclass
@@ -613,16 +615,16 @@ class Seq2SeqMoEModelOutput(ModelOutput):
613
615
  modules.
614
616
  """
615
617
 
616
- last_hidden_state: torch.FloatTensor | None = None
617
- past_key_values: EncoderDecoderCache | None = None
618
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
619
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
620
- decoder_router_logits: tuple[torch.FloatTensor] | None = None
621
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
622
- encoder_last_hidden_state: torch.FloatTensor | None = None
623
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
624
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
625
- encoder_router_logits: tuple[torch.FloatTensor] | None = None
618
+ last_hidden_state: Optional[torch.FloatTensor] = None
619
+ past_key_values: Optional[EncoderDecoderCache] = None
620
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
621
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
622
+ decoder_router_logits: Optional[tuple[torch.FloatTensor]] = None
623
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
624
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
625
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
626
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
627
+ encoder_router_logits: Optional[tuple[torch.FloatTensor]] = None
626
628
 
627
629
 
628
630
  @dataclass
@@ -648,10 +650,10 @@ class CausalLMOutput(ModelOutput):
648
650
  heads.
649
651
  """
650
652
 
651
- loss: torch.FloatTensor | None = None
652
- logits: torch.FloatTensor | None = None
653
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
654
- attentions: tuple[torch.FloatTensor, ...] | None = None
653
+ loss: Optional[torch.FloatTensor] = None
654
+ logits: Optional[torch.FloatTensor] = None
655
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
656
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
655
657
 
656
658
 
657
659
  @dataclass
@@ -682,11 +684,11 @@ class CausalLMOutputWithPast(ModelOutput):
682
684
  heads.
683
685
  """
684
686
 
685
- loss: torch.FloatTensor | None = None
686
- logits: torch.FloatTensor | None = None
687
- past_key_values: Cache | None = None
688
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
689
- attentions: tuple[torch.FloatTensor, ...] | None = None
687
+ loss: Optional[torch.FloatTensor] = None
688
+ logits: Optional[torch.FloatTensor] = None
689
+ past_key_values: Optional[Cache] = None
690
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
691
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
690
692
 
691
693
 
692
694
  @dataclass
@@ -723,12 +725,12 @@ class CausalLMOutputWithCrossAttentions(ModelOutput):
723
725
  `past_key_values` input) to speed up sequential decoding.
724
726
  """
725
727
 
726
- loss: torch.FloatTensor | None = None
727
- logits: torch.FloatTensor | None = None
728
- past_key_values: Cache | None = None
729
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
730
- attentions: tuple[torch.FloatTensor, ...] | None = None
731
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
728
+ loss: Optional[torch.FloatTensor] = None
729
+ logits: Optional[torch.FloatTensor] = None
730
+ past_key_values: Optional[Cache] = None
731
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
732
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
733
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
732
734
 
733
735
 
734
736
  @dataclass
@@ -759,11 +761,11 @@ class SequenceClassifierOutputWithPast(ModelOutput):
759
761
  heads.
760
762
  """
761
763
 
762
- loss: torch.FloatTensor | None = None
763
- logits: torch.FloatTensor | None = None
764
- past_key_values: Cache | None = None
765
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
766
- attentions: tuple[torch.FloatTensor, ...] | None = None
764
+ loss: Optional[torch.FloatTensor] = None
765
+ logits: Optional[torch.FloatTensor] = None
766
+ past_key_values: Optional[Cache] = None
767
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
768
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
767
769
 
768
770
 
769
771
  @dataclass
@@ -789,10 +791,10 @@ class MaskedLMOutput(ModelOutput):
789
791
  heads.
790
792
  """
791
793
 
792
- loss: torch.FloatTensor | None = None
793
- logits: torch.FloatTensor | None = None
794
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
795
- attentions: tuple[torch.FloatTensor, ...] | None = None
794
+ loss: Optional[torch.FloatTensor] = None
795
+ logits: Optional[torch.FloatTensor] = None
796
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
797
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
796
798
 
797
799
 
798
800
  @dataclass
@@ -842,15 +844,15 @@ class Seq2SeqLMOutput(ModelOutput):
842
844
  self-attention heads.
843
845
  """
844
846
 
845
- loss: torch.FloatTensor | None = None
846
- logits: torch.FloatTensor | None = None
847
- past_key_values: EncoderDecoderCache | None = None
848
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
849
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
850
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
851
- encoder_last_hidden_state: torch.FloatTensor | None = None
852
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
853
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
847
+ loss: Optional[torch.FloatTensor] = None
848
+ logits: Optional[torch.FloatTensor] = None
849
+ past_key_values: Optional[EncoderDecoderCache] = None
850
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
851
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
852
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
853
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
854
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
855
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
854
856
 
855
857
 
856
858
  @dataclass
@@ -909,21 +911,21 @@ class Seq2SeqMoEOutput(ModelOutput):
909
911
  models.
910
912
  """
911
913
 
912
- loss: torch.FloatTensor | None = None
913
- logits: torch.FloatTensor | None = None
914
- encoder_z_loss: torch.FloatTensor | None = None
915
- decoder_z_loss: torch.FloatTensor | None = None
916
- encoder_aux_loss: torch.FloatTensor | None = None
917
- decoder_aux_loss: torch.FloatTensor | None = None
918
- past_key_values: EncoderDecoderCache | None = None
919
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
920
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
921
- decoder_router_logits: tuple[torch.FloatTensor] | None = None
922
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
923
- encoder_last_hidden_state: torch.FloatTensor | None = None
924
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
925
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
926
- encoder_router_logits: tuple[torch.FloatTensor] | None = None
914
+ loss: Optional[torch.FloatTensor] = None
915
+ logits: Optional[torch.FloatTensor] = None
916
+ encoder_z_loss: Optional[torch.FloatTensor] = None
917
+ decoder_z_loss: Optional[torch.FloatTensor] = None
918
+ encoder_aux_loss: Optional[torch.FloatTensor] = None
919
+ decoder_aux_loss: Optional[torch.FloatTensor] = None
920
+ past_key_values: Optional[EncoderDecoderCache] = None
921
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
922
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
923
+ decoder_router_logits: Optional[tuple[torch.FloatTensor]] = None
924
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
925
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
926
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
927
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
928
+ encoder_router_logits: Optional[tuple[torch.FloatTensor]] = None
927
929
 
928
930
 
929
931
  @dataclass
@@ -950,10 +952,10 @@ class NextSentencePredictorOutput(ModelOutput):
950
952
  heads.
951
953
  """
952
954
 
953
- loss: torch.FloatTensor | None = None
954
- logits: torch.FloatTensor | None = None
955
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
956
- attentions: tuple[torch.FloatTensor, ...] | None = None
955
+ loss: Optional[torch.FloatTensor] = None
956
+ logits: Optional[torch.FloatTensor] = None
957
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
958
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
957
959
 
958
960
 
959
961
  @dataclass
@@ -979,10 +981,10 @@ class SequenceClassifierOutput(ModelOutput):
979
981
  heads.
980
982
  """
981
983
 
982
- loss: torch.FloatTensor | None = None
983
- logits: torch.FloatTensor | None = None
984
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
985
- attentions: tuple[torch.FloatTensor, ...] | None = None
984
+ loss: Optional[torch.FloatTensor] = None
985
+ logits: Optional[torch.FloatTensor] = None
986
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
987
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
986
988
 
987
989
 
988
990
  @dataclass
@@ -1032,15 +1034,15 @@ class Seq2SeqSequenceClassifierOutput(ModelOutput):
1032
1034
  self-attention heads.
1033
1035
  """
1034
1036
 
1035
- loss: torch.FloatTensor | None = None
1036
- logits: torch.FloatTensor | None = None
1037
- past_key_values: EncoderDecoderCache | None = None
1038
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1039
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1040
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
1041
- encoder_last_hidden_state: torch.FloatTensor | None = None
1042
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1043
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1037
+ loss: Optional[torch.FloatTensor] = None
1038
+ logits: Optional[torch.FloatTensor] = None
1039
+ past_key_values: Optional[EncoderDecoderCache] = None
1040
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1041
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1042
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1043
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
1044
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1045
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1044
1046
 
1045
1047
 
1046
1048
  @dataclass
@@ -1068,10 +1070,10 @@ class MultipleChoiceModelOutput(ModelOutput):
1068
1070
  heads.
1069
1071
  """
1070
1072
 
1071
- loss: torch.FloatTensor | None = None
1072
- logits: torch.FloatTensor | None = None
1073
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1074
- attentions: tuple[torch.FloatTensor, ...] | None = None
1073
+ loss: Optional[torch.FloatTensor] = None
1074
+ logits: Optional[torch.FloatTensor] = None
1075
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1076
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1075
1077
 
1076
1078
 
1077
1079
  @dataclass
@@ -1097,10 +1099,10 @@ class TokenClassifierOutput(ModelOutput):
1097
1099
  heads.
1098
1100
  """
1099
1101
 
1100
- loss: torch.FloatTensor | None = None
1101
- logits: torch.FloatTensor | None = None
1102
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1103
- attentions: tuple[torch.FloatTensor, ...] | None = None
1102
+ loss: Optional[torch.FloatTensor] = None
1103
+ logits: Optional[torch.FloatTensor] = None
1104
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1105
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1104
1106
 
1105
1107
 
1106
1108
  @dataclass
@@ -1128,11 +1130,11 @@ class QuestionAnsweringModelOutput(ModelOutput):
1128
1130
  heads.
1129
1131
  """
1130
1132
 
1131
- loss: torch.FloatTensor | None = None
1132
- start_logits: torch.FloatTensor | None = None
1133
- end_logits: torch.FloatTensor | None = None
1134
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1135
- attentions: tuple[torch.FloatTensor, ...] | None = None
1133
+ loss: Optional[torch.FloatTensor] = None
1134
+ start_logits: Optional[torch.FloatTensor] = None
1135
+ end_logits: Optional[torch.FloatTensor] = None
1136
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1137
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1136
1138
 
1137
1139
 
1138
1140
  @dataclass
@@ -1184,16 +1186,16 @@ class Seq2SeqQuestionAnsweringModelOutput(ModelOutput):
1184
1186
  self-attention heads.
1185
1187
  """
1186
1188
 
1187
- loss: torch.FloatTensor | None = None
1188
- start_logits: torch.FloatTensor | None = None
1189
- end_logits: torch.FloatTensor | None = None
1190
- past_key_values: EncoderDecoderCache | None = None
1191
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1192
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1193
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
1194
- encoder_last_hidden_state: torch.FloatTensor | None = None
1195
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1196
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1189
+ loss: Optional[torch.FloatTensor] = None
1190
+ start_logits: Optional[torch.FloatTensor] = None
1191
+ end_logits: Optional[torch.FloatTensor] = None
1192
+ past_key_values: Optional[EncoderDecoderCache] = None
1193
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1194
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1195
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1196
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
1197
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1198
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1197
1199
 
1198
1200
 
1199
1201
  @dataclass
@@ -1228,10 +1230,10 @@ class SemanticSegmenterOutput(ModelOutput):
1228
1230
  heads.
1229
1231
  """
1230
1232
 
1231
- loss: torch.FloatTensor | None = None
1232
- logits: torch.FloatTensor | None = None
1233
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1234
- attentions: tuple[torch.FloatTensor, ...] | None = None
1233
+ loss: Optional[torch.FloatTensor] = None
1234
+ logits: Optional[torch.FloatTensor] = None
1235
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1236
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1235
1237
 
1236
1238
 
1237
1239
  @dataclass
@@ -1256,10 +1258,10 @@ class ImageClassifierOutput(ModelOutput):
1256
1258
  heads.
1257
1259
  """
1258
1260
 
1259
- loss: torch.FloatTensor | None = None
1260
- logits: torch.FloatTensor | None = None
1261
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1262
- attentions: tuple[torch.FloatTensor, ...] | None = None
1261
+ loss: Optional[torch.FloatTensor] = None
1262
+ logits: Optional[torch.FloatTensor] = None
1263
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1264
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1263
1265
 
1264
1266
 
1265
1267
  @dataclass
@@ -1278,9 +1280,9 @@ class ImageClassifierOutputWithNoAttention(ModelOutput):
1278
1280
  called feature maps) of the model at the output of each stage.
1279
1281
  """
1280
1282
 
1281
- loss: torch.FloatTensor | None = None
1282
- logits: torch.FloatTensor | None = None
1283
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1283
+ loss: Optional[torch.FloatTensor] = None
1284
+ logits: Optional[torch.FloatTensor] = None
1285
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1284
1286
 
1285
1287
 
1286
1288
  @dataclass
@@ -1307,10 +1309,10 @@ class DepthEstimatorOutput(ModelOutput):
1307
1309
  heads.
1308
1310
  """
1309
1311
 
1310
- loss: torch.FloatTensor | None = None
1311
- predicted_depth: torch.FloatTensor | None = None
1312
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1313
- attentions: tuple[torch.FloatTensor, ...] | None = None
1312
+ loss: Optional[torch.FloatTensor] = None
1313
+ predicted_depth: Optional[torch.FloatTensor] = None
1314
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1315
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1314
1316
 
1315
1317
 
1316
1318
  @dataclass
@@ -1335,10 +1337,10 @@ class ImageSuperResolutionOutput(ModelOutput):
1335
1337
  heads.
1336
1338
  """
1337
1339
 
1338
- loss: torch.FloatTensor | None = None
1339
- reconstruction: torch.FloatTensor | None = None
1340
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1341
- attentions: tuple[torch.FloatTensor, ...] | None = None
1340
+ loss: Optional[torch.FloatTensor] = None
1341
+ reconstruction: Optional[torch.FloatTensor] = None
1342
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1343
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1342
1344
 
1343
1345
 
1344
1346
  @dataclass
@@ -1364,10 +1366,10 @@ class Wav2Vec2BaseModelOutput(ModelOutput):
1364
1366
  heads.
1365
1367
  """
1366
1368
 
1367
- last_hidden_state: torch.FloatTensor | None = None
1368
- extract_features: torch.FloatTensor | None = None
1369
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1370
- attentions: tuple[torch.FloatTensor, ...] | None = None
1369
+ last_hidden_state: Optional[torch.FloatTensor] = None
1370
+ extract_features: Optional[torch.FloatTensor] = None
1371
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1372
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1371
1373
 
1372
1374
 
1373
1375
  @dataclass
@@ -1395,11 +1397,11 @@ class XVectorOutput(ModelOutput):
1395
1397
  heads.
1396
1398
  """
1397
1399
 
1398
- loss: torch.FloatTensor | None = None
1399
- logits: torch.FloatTensor | None = None
1400
- embeddings: torch.FloatTensor | None = None
1401
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1402
- attentions: tuple[torch.FloatTensor, ...] | None = None
1400
+ loss: Optional[torch.FloatTensor] = None
1401
+ logits: Optional[torch.FloatTensor] = None
1402
+ embeddings: Optional[torch.FloatTensor] = None
1403
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1404
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1403
1405
 
1404
1406
 
1405
1407
  @dataclass
@@ -1424,9 +1426,9 @@ class BackboneOutput(ModelOutput):
1424
1426
  heads.
1425
1427
  """
1426
1428
 
1427
- feature_maps: tuple[torch.FloatTensor] | None = None
1428
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1429
- attentions: tuple[torch.FloatTensor, ...] | None = None
1429
+ feature_maps: Optional[tuple[torch.FloatTensor]] = None
1430
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1431
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1430
1432
 
1431
1433
 
1432
1434
  @dataclass
@@ -1459,11 +1461,11 @@ class BaseModelOutputWithPoolingAndProjection(ModelOutput):
1459
1461
  Text embeddings before the projection layer, used to mimic the last hidden state of the teacher encoder.
1460
1462
  """
1461
1463
 
1462
- last_hidden_state: torch.FloatTensor | None = None
1463
- pooler_output: torch.FloatTensor | None = None
1464
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1465
- attentions: tuple[torch.FloatTensor, ...] | None = None
1466
- projection_state: tuple[torch.FloatTensor] | None = None
1464
+ last_hidden_state: Optional[torch.FloatTensor] = None
1465
+ pooler_output: Optional[torch.FloatTensor] = None
1466
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1467
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1468
+ projection_state: Optional[tuple[torch.FloatTensor]] = None
1467
1469
 
1468
1470
 
1469
1471
  @dataclass
@@ -1513,15 +1515,15 @@ class Seq2SeqSpectrogramOutput(ModelOutput):
1513
1515
  self-attention heads.
1514
1516
  """
1515
1517
 
1516
- loss: torch.FloatTensor | None = None
1517
- spectrogram: torch.FloatTensor | None = None
1518
- past_key_values: EncoderDecoderCache | None = None
1519
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1520
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1521
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
1522
- encoder_last_hidden_state: torch.FloatTensor | None = None
1523
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1524
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1518
+ loss: Optional[torch.FloatTensor] = None
1519
+ spectrogram: Optional[torch.FloatTensor] = None
1520
+ past_key_values: Optional[EncoderDecoderCache] = None
1521
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1522
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1523
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1524
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
1525
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1526
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1525
1527
 
1526
1528
 
1527
1529
  @dataclass
@@ -1581,17 +1583,17 @@ class Seq2SeqTSModelOutput(ModelOutput):
1581
1583
  Static features of each time series' in a batch which are copied to the covariates at inference time.
1582
1584
  """
1583
1585
 
1584
- last_hidden_state: torch.FloatTensor | None = None
1585
- past_key_values: EncoderDecoderCache | None = None
1586
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1587
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1588
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
1589
- encoder_last_hidden_state: torch.FloatTensor | None = None
1590
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1591
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1592
- loc: torch.FloatTensor | None = None
1593
- scale: torch.FloatTensor | None = None
1594
- static_features: torch.FloatTensor | None = None
1586
+ last_hidden_state: Optional[torch.FloatTensor] = None
1587
+ past_key_values: Optional[EncoderDecoderCache] = None
1588
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1589
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1590
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1591
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
1592
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1593
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1594
+ loc: Optional[torch.FloatTensor] = None
1595
+ scale: Optional[torch.FloatTensor] = None
1596
+ static_features: Optional[torch.FloatTensor] = None
1595
1597
 
1596
1598
 
1597
1599
  @dataclass
@@ -1650,18 +1652,18 @@ class Seq2SeqTSPredictionOutput(ModelOutput):
1650
1652
  Static features of each time series' in a batch which are copied to the covariates at inference time.
1651
1653
  """
1652
1654
 
1653
- loss: torch.FloatTensor | None = None
1654
- params: tuple[torch.FloatTensor, ...] | None = None
1655
- past_key_values: EncoderDecoderCache | None = None
1656
- decoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1657
- decoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1658
- cross_attentions: tuple[torch.FloatTensor, ...] | None = None
1659
- encoder_last_hidden_state: torch.FloatTensor | None = None
1660
- encoder_hidden_states: tuple[torch.FloatTensor, ...] | None = None
1661
- encoder_attentions: tuple[torch.FloatTensor, ...] | None = None
1662
- loc: torch.FloatTensor | None = None
1663
- scale: torch.FloatTensor | None = None
1664
- static_features: torch.FloatTensor | None = None
1655
+ loss: Optional[torch.FloatTensor] = None
1656
+ params: Optional[tuple[torch.FloatTensor, ...]] = None
1657
+ past_key_values: Optional[EncoderDecoderCache] = None
1658
+ decoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1659
+ decoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1660
+ cross_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1661
+ encoder_last_hidden_state: Optional[torch.FloatTensor] = None
1662
+ encoder_hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1663
+ encoder_attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1664
+ loc: Optional[torch.FloatTensor] = None
1665
+ scale: Optional[torch.FloatTensor] = None
1666
+ static_features: Optional[torch.FloatTensor] = None
1665
1667
 
1666
1668
 
1667
1669
  @dataclass
@@ -1675,7 +1677,7 @@ class SampleTSPredictionOutput(ModelOutput):
1675
1677
  Sampled values from the chosen distribution.
1676
1678
  """
1677
1679
 
1678
- sequences: torch.FloatTensor | None = None
1680
+ sequences: Optional[torch.FloatTensor] = None
1679
1681
 
1680
1682
 
1681
1683
  @dataclass
@@ -1700,7 +1702,16 @@ class MaskedImageModelingOutput(ModelOutput):
1700
1702
  the self-attention heads.
1701
1703
  """
1702
1704
 
1703
- loss: torch.FloatTensor | None = None
1704
- reconstruction: torch.FloatTensor | None = None
1705
- hidden_states: tuple[torch.FloatTensor, ...] | None = None
1706
- attentions: tuple[torch.FloatTensor, ...] | None = None
1705
+ loss: Optional[torch.FloatTensor] = None
1706
+ reconstruction: Optional[torch.FloatTensor] = None
1707
+ hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1708
+ attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1709
+
1710
+ @property
1711
+ def logits(self):
1712
+ warnings.warn(
1713
+ "logits attribute is deprecated and will be removed in version 5 of Transformers."
1714
+ " Please use the reconstruction attribute to retrieve the final output instead.",
1715
+ FutureWarning,
1716
+ )
1717
+ return self.reconstruction