transformers 5.0.0rc1__py3-none-any.whl → 5.0.0rc3__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 (1584) hide show
  1. transformers/__init__.py +27 -27
  2. transformers/activations.py +1 -1
  3. transformers/audio_utils.py +32 -33
  4. transformers/cache_utils.py +32 -139
  5. transformers/cli/chat.py +3 -3
  6. transformers/cli/serve.py +2 -2
  7. transformers/cli/transformers.py +2 -1
  8. transformers/configuration_utils.py +143 -101
  9. transformers/conversion_mapping.py +73 -6
  10. transformers/convert_slow_tokenizer.py +3 -8
  11. transformers/core_model_loading.py +215 -50
  12. transformers/data/processors/glue.py +0 -1
  13. transformers/data/processors/utils.py +0 -1
  14. transformers/data/processors/xnli.py +0 -1
  15. transformers/dependency_versions_table.py +5 -5
  16. transformers/distributed/configuration_utils.py +1 -2
  17. transformers/dynamic_module_utils.py +23 -23
  18. transformers/feature_extraction_sequence_utils.py +19 -23
  19. transformers/feature_extraction_utils.py +63 -31
  20. transformers/generation/candidate_generator.py +80 -33
  21. transformers/generation/configuration_utils.py +186 -131
  22. transformers/generation/continuous_batching/__init__.py +0 -1
  23. transformers/generation/continuous_batching/cache.py +81 -24
  24. transformers/generation/continuous_batching/cache_manager.py +155 -45
  25. transformers/generation/continuous_batching/continuous_api.py +152 -84
  26. transformers/generation/continuous_batching/requests.py +51 -3
  27. transformers/generation/continuous_batching/scheduler.py +127 -52
  28. transformers/generation/logits_process.py +0 -128
  29. transformers/generation/stopping_criteria.py +1 -1
  30. transformers/generation/streamers.py +0 -1
  31. transformers/generation/utils.py +107 -119
  32. transformers/generation/watermarking.py +8 -6
  33. transformers/hf_argparser.py +9 -13
  34. transformers/hyperparameter_search.py +1 -2
  35. transformers/image_processing_base.py +11 -21
  36. transformers/image_processing_utils.py +11 -12
  37. transformers/image_processing_utils_fast.py +68 -57
  38. transformers/image_transforms.py +29 -29
  39. transformers/image_utils.py +30 -32
  40. transformers/initialization.py +37 -0
  41. transformers/integrations/__init__.py +12 -0
  42. transformers/integrations/accelerate.py +44 -111
  43. transformers/integrations/aqlm.py +3 -5
  44. transformers/integrations/awq.py +3 -8
  45. transformers/integrations/bitnet.py +5 -8
  46. transformers/integrations/bitsandbytes.py +16 -15
  47. transformers/integrations/deepspeed.py +19 -4
  48. transformers/integrations/eetq.py +3 -6
  49. transformers/integrations/fbgemm_fp8.py +2 -3
  50. transformers/integrations/finegrained_fp8.py +14 -23
  51. transformers/integrations/flash_attention.py +2 -2
  52. transformers/integrations/flex_attention.py +1 -1
  53. transformers/integrations/fp_quant.py +4 -6
  54. transformers/integrations/ggml.py +0 -1
  55. transformers/integrations/higgs.py +2 -5
  56. transformers/integrations/hub_kernels.py +23 -5
  57. transformers/integrations/integration_utils.py +37 -3
  58. transformers/integrations/mistral.py +12 -0
  59. transformers/integrations/moe.py +240 -0
  60. transformers/integrations/mxfp4.py +9 -16
  61. transformers/integrations/peft.py +5 -0
  62. transformers/integrations/quanto.py +5 -2
  63. transformers/integrations/quark.py +2 -4
  64. transformers/integrations/spqr.py +3 -5
  65. transformers/integrations/tensor_parallel.py +167 -221
  66. transformers/integrations/torchao.py +4 -6
  67. transformers/integrations/vptq.py +3 -5
  68. transformers/loss/loss_lw_detr.py +356 -0
  69. transformers/loss/loss_utils.py +2 -0
  70. transformers/masking_utils.py +47 -51
  71. transformers/model_debugging_utils.py +4 -5
  72. transformers/modelcard.py +14 -192
  73. transformers/modeling_attn_mask_utils.py +19 -19
  74. transformers/modeling_flash_attention_utils.py +27 -27
  75. transformers/modeling_gguf_pytorch_utils.py +71 -24
  76. transformers/modeling_layers.py +21 -22
  77. transformers/modeling_outputs.py +242 -253
  78. transformers/modeling_rope_utils.py +110 -113
  79. transformers/modeling_utils.py +633 -576
  80. transformers/models/__init__.py +23 -0
  81. transformers/models/afmoe/configuration_afmoe.py +26 -29
  82. transformers/models/afmoe/modeling_afmoe.py +37 -49
  83. transformers/models/afmoe/modular_afmoe.py +21 -31
  84. transformers/models/aimv2/configuration_aimv2.py +2 -5
  85. transformers/models/aimv2/modeling_aimv2.py +24 -21
  86. transformers/models/aimv2/modular_aimv2.py +11 -9
  87. transformers/models/albert/configuration_albert.py +0 -1
  88. transformers/models/albert/modeling_albert.py +70 -69
  89. transformers/models/albert/tokenization_albert.py +1 -4
  90. transformers/models/align/configuration_align.py +0 -1
  91. transformers/models/align/modeling_align.py +73 -68
  92. transformers/models/align/processing_align.py +2 -30
  93. transformers/models/altclip/configuration_altclip.py +0 -1
  94. transformers/models/altclip/modeling_altclip.py +83 -80
  95. transformers/models/altclip/processing_altclip.py +2 -15
  96. transformers/models/apertus/__init__.py +0 -1
  97. transformers/models/apertus/configuration_apertus.py +18 -21
  98. transformers/models/apertus/modeling_apertus.py +35 -36
  99. transformers/models/apertus/modular_apertus.py +32 -31
  100. transformers/models/arcee/configuration_arcee.py +20 -23
  101. transformers/models/arcee/modeling_arcee.py +32 -35
  102. transformers/models/arcee/modular_arcee.py +20 -23
  103. transformers/models/aria/configuration_aria.py +20 -23
  104. transformers/models/aria/image_processing_aria.py +25 -27
  105. transformers/models/aria/modeling_aria.py +71 -70
  106. transformers/models/aria/modular_aria.py +85 -88
  107. transformers/models/aria/processing_aria.py +28 -35
  108. transformers/models/audio_spectrogram_transformer/configuration_audio_spectrogram_transformer.py +0 -1
  109. transformers/models/audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.py +3 -6
  110. transformers/models/audio_spectrogram_transformer/modeling_audio_spectrogram_transformer.py +6 -8
  111. transformers/models/audioflamingo3/__init__.py +0 -1
  112. transformers/models/audioflamingo3/configuration_audioflamingo3.py +0 -1
  113. transformers/models/audioflamingo3/modeling_audioflamingo3.py +22 -23
  114. transformers/models/audioflamingo3/modular_audioflamingo3.py +12 -17
  115. transformers/models/audioflamingo3/processing_audioflamingo3.py +33 -30
  116. transformers/models/auto/auto_factory.py +5 -6
  117. transformers/models/auto/configuration_auto.py +53 -5
  118. transformers/models/auto/feature_extraction_auto.py +12 -10
  119. transformers/models/auto/image_processing_auto.py +17 -28
  120. transformers/models/auto/modeling_auto.py +38 -188
  121. transformers/models/auto/processing_auto.py +6 -1
  122. transformers/models/auto/tokenization_auto.py +147 -169
  123. transformers/models/auto/video_processing_auto.py +12 -10
  124. transformers/models/autoformer/configuration_autoformer.py +4 -7
  125. transformers/models/autoformer/modeling_autoformer.py +98 -100
  126. transformers/models/aya_vision/configuration_aya_vision.py +0 -1
  127. transformers/models/aya_vision/modeling_aya_vision.py +42 -40
  128. transformers/models/aya_vision/modular_aya_vision.py +26 -29
  129. transformers/models/aya_vision/processing_aya_vision.py +25 -53
  130. transformers/models/bamba/configuration_bamba.py +29 -32
  131. transformers/models/bamba/modeling_bamba.py +78 -83
  132. transformers/models/bamba/modular_bamba.py +68 -71
  133. transformers/models/bark/configuration_bark.py +4 -7
  134. transformers/models/bark/generation_configuration_bark.py +3 -5
  135. transformers/models/bark/modeling_bark.py +49 -55
  136. transformers/models/bark/processing_bark.py +19 -41
  137. transformers/models/bart/configuration_bart.py +0 -2
  138. transformers/models/bart/modeling_bart.py +122 -117
  139. transformers/models/barthez/tokenization_barthez.py +1 -4
  140. transformers/models/bartpho/tokenization_bartpho.py +6 -7
  141. transformers/models/beit/configuration_beit.py +0 -11
  142. transformers/models/beit/image_processing_beit.py +53 -56
  143. transformers/models/beit/image_processing_beit_fast.py +8 -10
  144. transformers/models/beit/modeling_beit.py +51 -53
  145. transformers/models/bert/configuration_bert.py +0 -1
  146. transformers/models/bert/modeling_bert.py +114 -122
  147. transformers/models/bert/tokenization_bert.py +2 -4
  148. transformers/models/bert/tokenization_bert_legacy.py +3 -5
  149. transformers/models/bert_generation/configuration_bert_generation.py +0 -1
  150. transformers/models/bert_generation/modeling_bert_generation.py +49 -49
  151. transformers/models/bert_generation/tokenization_bert_generation.py +2 -3
  152. transformers/models/bert_japanese/tokenization_bert_japanese.py +5 -6
  153. transformers/models/bertweet/tokenization_bertweet.py +1 -3
  154. transformers/models/big_bird/configuration_big_bird.py +0 -1
  155. transformers/models/big_bird/modeling_big_bird.py +110 -109
  156. transformers/models/big_bird/tokenization_big_bird.py +1 -4
  157. transformers/models/bigbird_pegasus/configuration_bigbird_pegasus.py +0 -1
  158. transformers/models/bigbird_pegasus/modeling_bigbird_pegasus.py +116 -111
  159. transformers/models/biogpt/configuration_biogpt.py +0 -1
  160. transformers/models/biogpt/modeling_biogpt.py +69 -71
  161. transformers/models/biogpt/modular_biogpt.py +59 -61
  162. transformers/models/biogpt/tokenization_biogpt.py +3 -5
  163. transformers/models/bit/configuration_bit.py +0 -1
  164. transformers/models/bit/image_processing_bit.py +21 -24
  165. transformers/models/bit/image_processing_bit_fast.py +0 -1
  166. transformers/models/bit/modeling_bit.py +14 -12
  167. transformers/models/bitnet/configuration_bitnet.py +18 -21
  168. transformers/models/bitnet/modeling_bitnet.py +32 -35
  169. transformers/models/bitnet/modular_bitnet.py +4 -6
  170. transformers/models/blenderbot/configuration_blenderbot.py +0 -1
  171. transformers/models/blenderbot/modeling_blenderbot.py +71 -95
  172. transformers/models/blenderbot/tokenization_blenderbot.py +6 -8
  173. transformers/models/blenderbot_small/configuration_blenderbot_small.py +0 -1
  174. transformers/models/blenderbot_small/modeling_blenderbot_small.py +73 -68
  175. transformers/models/blenderbot_small/tokenization_blenderbot_small.py +1 -3
  176. transformers/models/blip/configuration_blip.py +0 -1
  177. transformers/models/blip/image_processing_blip.py +17 -20
  178. transformers/models/blip/image_processing_blip_fast.py +0 -1
  179. transformers/models/blip/modeling_blip.py +62 -71
  180. transformers/models/blip/modeling_blip_text.py +71 -65
  181. transformers/models/blip/processing_blip.py +5 -36
  182. transformers/models/blip_2/configuration_blip_2.py +0 -1
  183. transformers/models/blip_2/modeling_blip_2.py +72 -71
  184. transformers/models/blip_2/processing_blip_2.py +8 -38
  185. transformers/models/bloom/configuration_bloom.py +0 -1
  186. transformers/models/bloom/modeling_bloom.py +71 -103
  187. transformers/models/blt/configuration_blt.py +71 -74
  188. transformers/models/blt/modeling_blt.py +235 -78
  189. transformers/models/blt/modular_blt.py +225 -62
  190. transformers/models/bridgetower/configuration_bridgetower.py +0 -1
  191. transformers/models/bridgetower/image_processing_bridgetower.py +34 -35
  192. transformers/models/bridgetower/image_processing_bridgetower_fast.py +7 -10
  193. transformers/models/bridgetower/modeling_bridgetower.py +113 -109
  194. transformers/models/bridgetower/processing_bridgetower.py +2 -16
  195. transformers/models/bros/configuration_bros.py +0 -1
  196. transformers/models/bros/modeling_bros.py +86 -80
  197. transformers/models/bros/processing_bros.py +2 -12
  198. transformers/models/byt5/tokenization_byt5.py +4 -6
  199. transformers/models/camembert/configuration_camembert.py +0 -1
  200. transformers/models/camembert/modeling_camembert.py +196 -195
  201. transformers/models/camembert/modular_camembert.py +51 -54
  202. transformers/models/camembert/tokenization_camembert.py +1 -4
  203. transformers/models/canine/configuration_canine.py +0 -1
  204. transformers/models/canine/modeling_canine.py +79 -75
  205. transformers/models/canine/tokenization_canine.py +2 -1
  206. transformers/models/chameleon/configuration_chameleon.py +24 -27
  207. transformers/models/chameleon/image_processing_chameleon.py +21 -24
  208. transformers/models/chameleon/image_processing_chameleon_fast.py +0 -1
  209. transformers/models/chameleon/modeling_chameleon.py +62 -60
  210. transformers/models/chameleon/processing_chameleon.py +16 -41
  211. transformers/models/chinese_clip/configuration_chinese_clip.py +0 -1
  212. transformers/models/chinese_clip/image_processing_chinese_clip.py +21 -24
  213. transformers/models/chinese_clip/image_processing_chinese_clip_fast.py +0 -1
  214. transformers/models/chinese_clip/modeling_chinese_clip.py +71 -69
  215. transformers/models/chinese_clip/processing_chinese_clip.py +2 -15
  216. transformers/models/clap/configuration_clap.py +0 -1
  217. transformers/models/clap/feature_extraction_clap.py +11 -12
  218. transformers/models/clap/modeling_clap.py +113 -104
  219. transformers/models/clap/processing_clap.py +2 -15
  220. transformers/models/clip/configuration_clip.py +0 -1
  221. transformers/models/clip/image_processing_clip.py +21 -24
  222. transformers/models/clip/image_processing_clip_fast.py +0 -1
  223. transformers/models/clip/modeling_clip.py +47 -46
  224. transformers/models/clip/processing_clip.py +2 -14
  225. transformers/models/clip/tokenization_clip.py +2 -5
  226. transformers/models/clipseg/configuration_clipseg.py +0 -1
  227. transformers/models/clipseg/modeling_clipseg.py +90 -87
  228. transformers/models/clipseg/processing_clipseg.py +8 -39
  229. transformers/models/clvp/configuration_clvp.py +1 -3
  230. transformers/models/clvp/feature_extraction_clvp.py +7 -10
  231. transformers/models/clvp/modeling_clvp.py +133 -118
  232. transformers/models/clvp/number_normalizer.py +1 -2
  233. transformers/models/clvp/processing_clvp.py +3 -20
  234. transformers/models/clvp/tokenization_clvp.py +0 -1
  235. transformers/models/code_llama/tokenization_code_llama.py +4 -7
  236. transformers/models/codegen/configuration_codegen.py +0 -1
  237. transformers/models/codegen/modeling_codegen.py +61 -52
  238. transformers/models/codegen/tokenization_codegen.py +5 -6
  239. transformers/models/cohere/configuration_cohere.py +20 -23
  240. transformers/models/cohere/modeling_cohere.py +36 -39
  241. transformers/models/cohere/modular_cohere.py +24 -28
  242. transformers/models/cohere/tokenization_cohere.py +5 -6
  243. transformers/models/cohere2/configuration_cohere2.py +21 -24
  244. transformers/models/cohere2/modeling_cohere2.py +35 -38
  245. transformers/models/cohere2/modular_cohere2.py +39 -41
  246. transformers/models/cohere2_vision/image_processing_cohere2_vision_fast.py +6 -8
  247. transformers/models/cohere2_vision/modeling_cohere2_vision.py +35 -33
  248. transformers/models/cohere2_vision/modular_cohere2_vision.py +21 -23
  249. transformers/models/cohere2_vision/processing_cohere2_vision.py +6 -36
  250. transformers/models/colpali/configuration_colpali.py +0 -1
  251. transformers/models/colpali/modeling_colpali.py +14 -16
  252. transformers/models/colpali/modular_colpali.py +11 -51
  253. transformers/models/colpali/processing_colpali.py +14 -52
  254. transformers/models/colqwen2/modeling_colqwen2.py +20 -22
  255. transformers/models/colqwen2/modular_colqwen2.py +29 -68
  256. transformers/models/colqwen2/processing_colqwen2.py +16 -52
  257. transformers/models/conditional_detr/configuration_conditional_detr.py +1 -2
  258. transformers/models/conditional_detr/image_processing_conditional_detr.py +64 -66
  259. transformers/models/conditional_detr/image_processing_conditional_detr_fast.py +22 -22
  260. transformers/models/conditional_detr/modeling_conditional_detr.py +82 -81
  261. transformers/models/conditional_detr/modular_conditional_detr.py +1 -3
  262. transformers/models/convbert/configuration_convbert.py +0 -1
  263. transformers/models/convbert/modeling_convbert.py +88 -87
  264. transformers/models/convbert/tokenization_convbert.py +0 -1
  265. transformers/models/convnext/configuration_convnext.py +0 -1
  266. transformers/models/convnext/image_processing_convnext.py +20 -23
  267. transformers/models/convnext/image_processing_convnext_fast.py +14 -19
  268. transformers/models/convnext/modeling_convnext.py +5 -8
  269. transformers/models/convnextv2/configuration_convnextv2.py +0 -1
  270. transformers/models/convnextv2/modeling_convnextv2.py +5 -8
  271. transformers/models/cpm/tokenization_cpm.py +6 -7
  272. transformers/models/cpm/tokenization_cpm_fast.py +3 -5
  273. transformers/models/cpmant/configuration_cpmant.py +0 -1
  274. transformers/models/cpmant/modeling_cpmant.py +38 -40
  275. transformers/models/cpmant/tokenization_cpmant.py +1 -3
  276. transformers/models/csm/configuration_csm.py +49 -51
  277. transformers/models/csm/generation_csm.py +31 -35
  278. transformers/models/csm/modeling_csm.py +81 -82
  279. transformers/models/csm/modular_csm.py +58 -58
  280. transformers/models/csm/processing_csm.py +25 -68
  281. transformers/models/ctrl/configuration_ctrl.py +0 -1
  282. transformers/models/ctrl/modeling_ctrl.py +52 -43
  283. transformers/models/ctrl/tokenization_ctrl.py +0 -1
  284. transformers/models/cvt/configuration_cvt.py +0 -1
  285. transformers/models/cvt/modeling_cvt.py +18 -16
  286. transformers/models/cwm/__init__.py +0 -1
  287. transformers/models/cwm/configuration_cwm.py +3 -5
  288. transformers/models/cwm/modeling_cwm.py +33 -35
  289. transformers/models/cwm/modular_cwm.py +10 -12
  290. transformers/models/d_fine/configuration_d_fine.py +3 -5
  291. transformers/models/d_fine/modeling_d_fine.py +127 -121
  292. transformers/models/d_fine/modular_d_fine.py +23 -13
  293. transformers/models/dab_detr/configuration_dab_detr.py +2 -3
  294. transformers/models/dab_detr/modeling_dab_detr.py +69 -71
  295. transformers/models/dac/configuration_dac.py +0 -1
  296. transformers/models/dac/feature_extraction_dac.py +6 -9
  297. transformers/models/dac/modeling_dac.py +21 -23
  298. transformers/models/data2vec/configuration_data2vec_audio.py +0 -1
  299. transformers/models/data2vec/configuration_data2vec_text.py +0 -1
  300. transformers/models/data2vec/configuration_data2vec_vision.py +0 -1
  301. transformers/models/data2vec/modeling_data2vec_audio.py +52 -56
  302. transformers/models/data2vec/modeling_data2vec_text.py +98 -93
  303. transformers/models/data2vec/modeling_data2vec_vision.py +41 -42
  304. transformers/models/data2vec/modular_data2vec_audio.py +6 -1
  305. transformers/models/data2vec/modular_data2vec_text.py +58 -54
  306. transformers/models/dbrx/configuration_dbrx.py +27 -20
  307. transformers/models/dbrx/modeling_dbrx.py +40 -43
  308. transformers/models/dbrx/modular_dbrx.py +31 -33
  309. transformers/models/deberta/configuration_deberta.py +0 -1
  310. transformers/models/deberta/modeling_deberta.py +59 -60
  311. transformers/models/deberta/tokenization_deberta.py +2 -5
  312. transformers/models/deberta_v2/configuration_deberta_v2.py +0 -1
  313. transformers/models/deberta_v2/modeling_deberta_v2.py +65 -65
  314. transformers/models/deberta_v2/tokenization_deberta_v2.py +1 -4
  315. transformers/models/decision_transformer/configuration_decision_transformer.py +0 -1
  316. transformers/models/decision_transformer/modeling_decision_transformer.py +56 -55
  317. transformers/models/deepseek_v2/configuration_deepseek_v2.py +34 -37
  318. transformers/models/deepseek_v2/modeling_deepseek_v2.py +39 -37
  319. transformers/models/deepseek_v2/modular_deepseek_v2.py +44 -44
  320. transformers/models/deepseek_v3/configuration_deepseek_v3.py +35 -38
  321. transformers/models/deepseek_v3/modeling_deepseek_v3.py +40 -38
  322. transformers/models/deepseek_v3/modular_deepseek_v3.py +10 -7
  323. transformers/models/deepseek_vl/configuration_deepseek_vl.py +2 -3
  324. transformers/models/deepseek_vl/image_processing_deepseek_vl.py +25 -26
  325. transformers/models/deepseek_vl/image_processing_deepseek_vl_fast.py +7 -7
  326. transformers/models/deepseek_vl/modeling_deepseek_vl.py +40 -36
  327. transformers/models/deepseek_vl/modular_deepseek_vl.py +14 -43
  328. transformers/models/deepseek_vl/processing_deepseek_vl.py +10 -41
  329. transformers/models/deepseek_vl_hybrid/configuration_deepseek_vl_hybrid.py +3 -5
  330. transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py +35 -35
  331. transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py +16 -20
  332. transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py +42 -38
  333. transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py +80 -99
  334. transformers/models/deepseek_vl_hybrid/processing_deepseek_vl_hybrid.py +12 -44
  335. transformers/models/deformable_detr/configuration_deformable_detr.py +2 -3
  336. transformers/models/deformable_detr/image_processing_deformable_detr.py +59 -61
  337. transformers/models/deformable_detr/image_processing_deformable_detr_fast.py +17 -17
  338. transformers/models/deformable_detr/modeling_deformable_detr.py +67 -68
  339. transformers/models/deformable_detr/modular_deformable_detr.py +1 -3
  340. transformers/models/deit/configuration_deit.py +0 -1
  341. transformers/models/deit/image_processing_deit.py +18 -21
  342. transformers/models/deit/image_processing_deit_fast.py +0 -1
  343. transformers/models/deit/modeling_deit.py +16 -18
  344. transformers/models/depth_anything/configuration_depth_anything.py +2 -4
  345. transformers/models/depth_anything/modeling_depth_anything.py +5 -8
  346. transformers/models/depth_pro/configuration_depth_pro.py +0 -1
  347. transformers/models/depth_pro/image_processing_depth_pro.py +22 -23
  348. transformers/models/depth_pro/image_processing_depth_pro_fast.py +6 -8
  349. transformers/models/depth_pro/modeling_depth_pro.py +21 -23
  350. transformers/models/detr/configuration_detr.py +1 -2
  351. transformers/models/detr/image_processing_detr.py +64 -66
  352. transformers/models/detr/image_processing_detr_fast.py +22 -23
  353. transformers/models/detr/modeling_detr.py +78 -73
  354. transformers/models/dia/configuration_dia.py +5 -8
  355. transformers/models/dia/feature_extraction_dia.py +6 -9
  356. transformers/models/dia/generation_dia.py +42 -45
  357. transformers/models/dia/modeling_dia.py +73 -65
  358. transformers/models/dia/modular_dia.py +63 -54
  359. transformers/models/dia/processing_dia.py +39 -29
  360. transformers/models/dia/tokenization_dia.py +3 -6
  361. transformers/models/diffllama/configuration_diffllama.py +20 -23
  362. transformers/models/diffllama/modeling_diffllama.py +44 -47
  363. transformers/models/diffllama/modular_diffllama.py +17 -19
  364. transformers/models/dinat/configuration_dinat.py +0 -1
  365. transformers/models/dinat/modeling_dinat.py +40 -42
  366. transformers/models/dinov2/configuration_dinov2.py +0 -1
  367. transformers/models/dinov2/modeling_dinov2.py +11 -13
  368. transformers/models/dinov2_with_registers/configuration_dinov2_with_registers.py +1 -1
  369. transformers/models/dinov2_with_registers/modeling_dinov2_with_registers.py +12 -13
  370. transformers/models/dinov2_with_registers/modular_dinov2_with_registers.py +5 -7
  371. transformers/models/dinov3_convnext/configuration_dinov3_convnext.py +4 -7
  372. transformers/models/dinov3_convnext/modeling_dinov3_convnext.py +3 -6
  373. transformers/models/dinov3_vit/configuration_dinov3_vit.py +5 -8
  374. transformers/models/dinov3_vit/image_processing_dinov3_vit_fast.py +5 -7
  375. transformers/models/dinov3_vit/modeling_dinov3_vit.py +17 -16
  376. transformers/models/dinov3_vit/modular_dinov3_vit.py +14 -13
  377. transformers/models/distilbert/configuration_distilbert.py +0 -1
  378. transformers/models/distilbert/modeling_distilbert.py +55 -55
  379. transformers/models/distilbert/tokenization_distilbert.py +0 -1
  380. transformers/models/doge/__init__.py +0 -1
  381. transformers/models/doge/configuration_doge.py +25 -28
  382. transformers/models/doge/modeling_doge.py +43 -46
  383. transformers/models/doge/modular_doge.py +57 -58
  384. transformers/models/donut/configuration_donut_swin.py +0 -1
  385. transformers/models/donut/image_processing_donut.py +26 -29
  386. transformers/models/donut/image_processing_donut_fast.py +5 -11
  387. transformers/models/donut/modeling_donut_swin.py +60 -58
  388. transformers/models/donut/processing_donut.py +5 -26
  389. transformers/models/dots1/configuration_dots1.py +27 -29
  390. transformers/models/dots1/modeling_dots1.py +45 -39
  391. transformers/models/dots1/modular_dots1.py +0 -1
  392. transformers/models/dpr/configuration_dpr.py +0 -1
  393. transformers/models/dpr/modeling_dpr.py +37 -39
  394. transformers/models/dpr/tokenization_dpr.py +7 -9
  395. transformers/models/dpr/tokenization_dpr_fast.py +7 -9
  396. transformers/models/dpt/configuration_dpt.py +1 -2
  397. transformers/models/dpt/image_processing_dpt.py +65 -66
  398. transformers/models/dpt/image_processing_dpt_fast.py +14 -16
  399. transformers/models/dpt/modeling_dpt.py +19 -21
  400. transformers/models/dpt/modular_dpt.py +11 -13
  401. transformers/models/edgetam/configuration_edgetam.py +1 -2
  402. transformers/models/edgetam/modeling_edgetam.py +44 -43
  403. transformers/models/edgetam/modular_edgetam.py +17 -20
  404. transformers/models/edgetam_video/__init__.py +0 -1
  405. transformers/models/edgetam_video/configuration_edgetam_video.py +0 -1
  406. transformers/models/edgetam_video/modeling_edgetam_video.py +131 -120
  407. transformers/models/edgetam_video/modular_edgetam_video.py +29 -37
  408. transformers/models/efficientloftr/configuration_efficientloftr.py +4 -5
  409. transformers/models/efficientloftr/image_processing_efficientloftr.py +14 -16
  410. transformers/models/efficientloftr/image_processing_efficientloftr_fast.py +5 -6
  411. transformers/models/efficientloftr/modeling_efficientloftr.py +41 -30
  412. transformers/models/efficientloftr/modular_efficientloftr.py +1 -3
  413. transformers/models/efficientnet/configuration_efficientnet.py +0 -1
  414. transformers/models/efficientnet/image_processing_efficientnet.py +28 -32
  415. transformers/models/efficientnet/image_processing_efficientnet_fast.py +15 -17
  416. transformers/models/efficientnet/modeling_efficientnet.py +17 -15
  417. transformers/models/electra/configuration_electra.py +0 -1
  418. transformers/models/electra/modeling_electra.py +108 -103
  419. transformers/models/emu3/configuration_emu3.py +5 -7
  420. transformers/models/emu3/image_processing_emu3.py +44 -39
  421. transformers/models/emu3/modeling_emu3.py +67 -64
  422. transformers/models/emu3/modular_emu3.py +39 -35
  423. transformers/models/emu3/processing_emu3.py +18 -43
  424. transformers/models/encodec/configuration_encodec.py +2 -4
  425. transformers/models/encodec/feature_extraction_encodec.py +10 -13
  426. transformers/models/encodec/modeling_encodec.py +39 -29
  427. transformers/models/encoder_decoder/configuration_encoder_decoder.py +0 -1
  428. transformers/models/encoder_decoder/modeling_encoder_decoder.py +17 -19
  429. transformers/models/eomt/configuration_eomt.py +0 -1
  430. transformers/models/eomt/image_processing_eomt.py +53 -55
  431. transformers/models/eomt/image_processing_eomt_fast.py +59 -28
  432. transformers/models/eomt/modeling_eomt.py +23 -18
  433. transformers/models/eomt/modular_eomt.py +18 -13
  434. transformers/models/ernie/configuration_ernie.py +0 -1
  435. transformers/models/ernie/modeling_ernie.py +127 -132
  436. transformers/models/ernie/modular_ernie.py +97 -103
  437. transformers/models/ernie4_5/configuration_ernie4_5.py +18 -20
  438. transformers/models/ernie4_5/modeling_ernie4_5.py +32 -34
  439. transformers/models/ernie4_5/modular_ernie4_5.py +1 -3
  440. transformers/models/ernie4_5_moe/configuration_ernie4_5_moe.py +27 -29
  441. transformers/models/ernie4_5_moe/modeling_ernie4_5_moe.py +52 -51
  442. transformers/models/ernie4_5_moe/modular_ernie4_5_moe.py +16 -44
  443. transformers/models/ernie4_5_vl_moe/__init__.py +31 -0
  444. transformers/models/ernie4_5_vl_moe/configuration_ernie4_5_vl_moe.py +329 -0
  445. transformers/models/ernie4_5_vl_moe/image_processing_ernie4_5_vl_moe.py +455 -0
  446. transformers/models/ernie4_5_vl_moe/image_processing_ernie4_5_vl_moe_fast.py +231 -0
  447. transformers/models/ernie4_5_vl_moe/modeling_ernie4_5_vl_moe.py +1895 -0
  448. transformers/models/ernie4_5_vl_moe/modular_ernie4_5_vl_moe.py +1901 -0
  449. transformers/models/ernie4_5_vl_moe/processing_ernie4_5_vl_moe.py +249 -0
  450. transformers/models/ernie4_5_vl_moe/video_processing_ernie4_5_vl_moe.py +593 -0
  451. transformers/models/esm/configuration_esm.py +2 -4
  452. transformers/models/esm/modeling_esm.py +38 -34
  453. transformers/models/esm/modeling_esmfold.py +48 -45
  454. transformers/models/esm/openfold_utils/chunk_utils.py +6 -6
  455. transformers/models/esm/openfold_utils/loss.py +1 -2
  456. transformers/models/esm/openfold_utils/protein.py +13 -13
  457. transformers/models/esm/openfold_utils/tensor_utils.py +6 -6
  458. transformers/models/esm/tokenization_esm.py +2 -4
  459. transformers/models/evolla/configuration_evolla.py +29 -32
  460. transformers/models/evolla/modeling_evolla.py +67 -62
  461. transformers/models/evolla/modular_evolla.py +53 -47
  462. transformers/models/evolla/processing_evolla.py +23 -35
  463. transformers/models/exaone4/configuration_exaone4.py +19 -22
  464. transformers/models/exaone4/modeling_exaone4.py +33 -36
  465. transformers/models/exaone4/modular_exaone4.py +40 -42
  466. transformers/models/falcon/configuration_falcon.py +22 -25
  467. transformers/models/falcon/modeling_falcon.py +75 -78
  468. transformers/models/falcon_h1/configuration_falcon_h1.py +40 -43
  469. transformers/models/falcon_h1/modeling_falcon_h1.py +80 -78
  470. transformers/models/falcon_h1/modular_falcon_h1.py +54 -50
  471. transformers/models/falcon_mamba/configuration_falcon_mamba.py +0 -1
  472. transformers/models/falcon_mamba/modeling_falcon_mamba.py +50 -47
  473. transformers/models/falcon_mamba/modular_falcon_mamba.py +16 -14
  474. transformers/models/fast_vlm/configuration_fast_vlm.py +1 -0
  475. transformers/models/fast_vlm/modeling_fast_vlm.py +43 -39
  476. transformers/models/fast_vlm/modular_fast_vlm.py +2 -3
  477. transformers/models/fastspeech2_conformer/configuration_fastspeech2_conformer.py +2 -5
  478. transformers/models/fastspeech2_conformer/modeling_fastspeech2_conformer.py +68 -57
  479. transformers/models/fastspeech2_conformer/tokenization_fastspeech2_conformer.py +2 -3
  480. transformers/models/flaubert/configuration_flaubert.py +0 -1
  481. transformers/models/flaubert/modeling_flaubert.py +138 -143
  482. transformers/models/flaubert/tokenization_flaubert.py +3 -5
  483. transformers/models/flava/configuration_flava.py +5 -6
  484. transformers/models/flava/image_processing_flava.py +66 -67
  485. transformers/models/flava/image_processing_flava_fast.py +42 -45
  486. transformers/models/flava/modeling_flava.py +111 -107
  487. transformers/models/flava/processing_flava.py +2 -12
  488. transformers/models/flex_olmo/__init__.py +0 -1
  489. transformers/models/flex_olmo/configuration_flex_olmo.py +23 -25
  490. transformers/models/flex_olmo/modeling_flex_olmo.py +44 -43
  491. transformers/models/flex_olmo/modular_flex_olmo.py +35 -37
  492. transformers/models/florence2/configuration_florence2.py +0 -1
  493. transformers/models/florence2/modeling_florence2.py +59 -43
  494. transformers/models/florence2/modular_florence2.py +65 -81
  495. transformers/models/florence2/processing_florence2.py +18 -47
  496. transformers/models/fnet/configuration_fnet.py +0 -1
  497. transformers/models/fnet/modeling_fnet.py +76 -80
  498. transformers/models/fnet/tokenization_fnet.py +0 -1
  499. transformers/models/focalnet/configuration_focalnet.py +0 -1
  500. transformers/models/focalnet/modeling_focalnet.py +39 -41
  501. transformers/models/fsmt/configuration_fsmt.py +0 -1
  502. transformers/models/fsmt/modeling_fsmt.py +47 -48
  503. transformers/models/fsmt/tokenization_fsmt.py +3 -5
  504. transformers/models/funnel/configuration_funnel.py +0 -1
  505. transformers/models/funnel/modeling_funnel.py +91 -93
  506. transformers/models/funnel/tokenization_funnel.py +2 -5
  507. transformers/models/fuyu/configuration_fuyu.py +23 -26
  508. transformers/models/fuyu/image_processing_fuyu.py +29 -31
  509. transformers/models/fuyu/image_processing_fuyu_fast.py +12 -13
  510. transformers/models/fuyu/modeling_fuyu.py +29 -30
  511. transformers/models/fuyu/processing_fuyu.py +23 -34
  512. transformers/models/gemma/configuration_gemma.py +20 -23
  513. transformers/models/gemma/modeling_gemma.py +42 -46
  514. transformers/models/gemma/modular_gemma.py +37 -40
  515. transformers/models/gemma/tokenization_gemma.py +3 -6
  516. transformers/models/gemma2/configuration_gemma2.py +25 -28
  517. transformers/models/gemma2/modeling_gemma2.py +35 -38
  518. transformers/models/gemma2/modular_gemma2.py +56 -58
  519. transformers/models/gemma3/configuration_gemma3.py +28 -29
  520. transformers/models/gemma3/image_processing_gemma3.py +29 -31
  521. transformers/models/gemma3/image_processing_gemma3_fast.py +9 -11
  522. transformers/models/gemma3/modeling_gemma3.py +112 -94
  523. transformers/models/gemma3/modular_gemma3.py +110 -91
  524. transformers/models/gemma3/processing_gemma3.py +5 -5
  525. transformers/models/gemma3n/configuration_gemma3n.py +12 -10
  526. transformers/models/gemma3n/feature_extraction_gemma3n.py +9 -11
  527. transformers/models/gemma3n/modeling_gemma3n.py +127 -98
  528. transformers/models/gemma3n/modular_gemma3n.py +117 -84
  529. transformers/models/gemma3n/processing_gemma3n.py +12 -26
  530. transformers/models/git/configuration_git.py +0 -1
  531. transformers/models/git/modeling_git.py +250 -197
  532. transformers/models/git/processing_git.py +2 -14
  533. transformers/models/glm/configuration_glm.py +19 -21
  534. transformers/models/glm/modeling_glm.py +33 -36
  535. transformers/models/glm/modular_glm.py +4 -7
  536. transformers/models/glm4/configuration_glm4.py +19 -21
  537. transformers/models/glm4/modeling_glm4.py +36 -38
  538. transformers/models/glm4/modular_glm4.py +8 -10
  539. transformers/models/glm46v/configuration_glm46v.py +0 -1
  540. transformers/models/glm46v/image_processing_glm46v.py +35 -40
  541. transformers/models/glm46v/image_processing_glm46v_fast.py +7 -7
  542. transformers/models/glm46v/modeling_glm46v.py +54 -52
  543. transformers/models/glm46v/modular_glm46v.py +4 -3
  544. transformers/models/glm46v/processing_glm46v.py +7 -41
  545. transformers/models/glm46v/video_processing_glm46v.py +9 -11
  546. transformers/models/glm4_moe/configuration_glm4_moe.py +25 -28
  547. transformers/models/glm4_moe/modeling_glm4_moe.py +41 -40
  548. transformers/models/glm4_moe/modular_glm4_moe.py +27 -30
  549. transformers/models/glm4_moe_lite/__init__.py +28 -0
  550. transformers/models/glm4_moe_lite/configuration_glm4_moe_lite.py +235 -0
  551. transformers/models/glm4_moe_lite/modeling_glm4_moe_lite.py +740 -0
  552. transformers/models/glm4_moe_lite/modular_glm4_moe_lite.py +304 -0
  553. transformers/models/glm4v/configuration_glm4v.py +14 -17
  554. transformers/models/glm4v/image_processing_glm4v.py +34 -40
  555. transformers/models/glm4v/image_processing_glm4v_fast.py +6 -7
  556. transformers/models/glm4v/modeling_glm4v.py +148 -156
  557. transformers/models/glm4v/modular_glm4v.py +142 -185
  558. transformers/models/glm4v/processing_glm4v.py +7 -41
  559. transformers/models/glm4v/video_processing_glm4v.py +9 -11
  560. transformers/models/glm4v_moe/configuration_glm4v_moe.py +119 -122
  561. transformers/models/glm4v_moe/modeling_glm4v_moe.py +275 -319
  562. transformers/models/glm4v_moe/modular_glm4v_moe.py +66 -163
  563. transformers/models/glm_image/__init__.py +31 -0
  564. transformers/models/glm_image/configuration_glm_image.py +352 -0
  565. transformers/models/glm_image/image_processing_glm_image.py +503 -0
  566. transformers/models/glm_image/image_processing_glm_image_fast.py +296 -0
  567. transformers/models/glm_image/modeling_glm_image.py +1590 -0
  568. transformers/models/glm_image/modular_glm_image.py +1480 -0
  569. transformers/models/glm_image/processing_glm_image.py +217 -0
  570. transformers/models/glmasr/__init__.py +29 -0
  571. transformers/models/glmasr/configuration_glmasr.py +196 -0
  572. transformers/models/glmasr/modeling_glmasr.py +511 -0
  573. transformers/models/glmasr/modular_glmasr.py +431 -0
  574. transformers/models/glmasr/processing_glmasr.py +331 -0
  575. transformers/models/glpn/configuration_glpn.py +0 -1
  576. transformers/models/glpn/image_processing_glpn.py +11 -12
  577. transformers/models/glpn/image_processing_glpn_fast.py +8 -10
  578. transformers/models/glpn/modeling_glpn.py +10 -12
  579. transformers/models/got_ocr2/configuration_got_ocr2.py +5 -8
  580. transformers/models/got_ocr2/image_processing_got_ocr2.py +22 -24
  581. transformers/models/got_ocr2/image_processing_got_ocr2_fast.py +6 -8
  582. transformers/models/got_ocr2/modeling_got_ocr2.py +48 -45
  583. transformers/models/got_ocr2/modular_got_ocr2.py +31 -34
  584. transformers/models/got_ocr2/processing_got_ocr2.py +42 -63
  585. transformers/models/gpt2/configuration_gpt2.py +0 -1
  586. transformers/models/gpt2/modeling_gpt2.py +114 -113
  587. transformers/models/gpt2/tokenization_gpt2.py +6 -9
  588. transformers/models/gpt_bigcode/configuration_gpt_bigcode.py +0 -1
  589. transformers/models/gpt_bigcode/modeling_gpt_bigcode.py +76 -88
  590. transformers/models/gpt_neo/configuration_gpt_neo.py +0 -1
  591. transformers/models/gpt_neo/modeling_gpt_neo.py +77 -66
  592. transformers/models/gpt_neox/configuration_gpt_neox.py +19 -22
  593. transformers/models/gpt_neox/modeling_gpt_neox.py +71 -73
  594. transformers/models/gpt_neox/modular_gpt_neox.py +64 -66
  595. transformers/models/gpt_neox/tokenization_gpt_neox.py +2 -5
  596. transformers/models/gpt_neox_japanese/configuration_gpt_neox_japanese.py +15 -18
  597. transformers/models/gpt_neox_japanese/modeling_gpt_neox_japanese.py +42 -45
  598. transformers/models/gpt_neox_japanese/tokenization_gpt_neox_japanese.py +1 -3
  599. transformers/models/gpt_oss/configuration_gpt_oss.py +38 -24
  600. transformers/models/gpt_oss/modeling_gpt_oss.py +40 -44
  601. transformers/models/gpt_oss/modular_gpt_oss.py +22 -26
  602. transformers/models/gpt_sw3/tokenization_gpt_sw3.py +4 -4
  603. transformers/models/gptj/configuration_gptj.py +0 -1
  604. transformers/models/gptj/modeling_gptj.py +96 -86
  605. transformers/models/granite/configuration_granite.py +23 -26
  606. transformers/models/granite/modeling_granite.py +40 -42
  607. transformers/models/granite/modular_granite.py +29 -31
  608. transformers/models/granite_speech/configuration_granite_speech.py +0 -1
  609. transformers/models/granite_speech/feature_extraction_granite_speech.py +1 -3
  610. transformers/models/granite_speech/modeling_granite_speech.py +36 -24
  611. transformers/models/granite_speech/processing_granite_speech.py +11 -4
  612. transformers/models/granitemoe/configuration_granitemoe.py +26 -29
  613. transformers/models/granitemoe/modeling_granitemoe.py +37 -40
  614. transformers/models/granitemoe/modular_granitemoe.py +22 -25
  615. transformers/models/granitemoehybrid/__init__.py +0 -1
  616. transformers/models/granitemoehybrid/configuration_granitemoehybrid.py +41 -40
  617. transformers/models/granitemoehybrid/modeling_granitemoehybrid.py +92 -86
  618. transformers/models/granitemoehybrid/modular_granitemoehybrid.py +29 -21
  619. transformers/models/granitemoeshared/configuration_granitemoeshared.py +27 -30
  620. transformers/models/granitemoeshared/modeling_granitemoeshared.py +50 -55
  621. transformers/models/granitemoeshared/modular_granitemoeshared.py +19 -21
  622. transformers/models/grounding_dino/configuration_grounding_dino.py +2 -4
  623. transformers/models/grounding_dino/image_processing_grounding_dino.py +60 -62
  624. transformers/models/grounding_dino/image_processing_grounding_dino_fast.py +17 -18
  625. transformers/models/grounding_dino/modeling_grounding_dino.py +95 -97
  626. transformers/models/grounding_dino/modular_grounding_dino.py +2 -3
  627. transformers/models/grounding_dino/processing_grounding_dino.py +10 -38
  628. transformers/models/groupvit/configuration_groupvit.py +0 -1
  629. transformers/models/groupvit/modeling_groupvit.py +75 -71
  630. transformers/models/helium/configuration_helium.py +20 -22
  631. transformers/models/helium/modeling_helium.py +34 -37
  632. transformers/models/helium/modular_helium.py +3 -7
  633. transformers/models/herbert/tokenization_herbert.py +4 -6
  634. transformers/models/hgnet_v2/configuration_hgnet_v2.py +0 -1
  635. transformers/models/hgnet_v2/modeling_hgnet_v2.py +16 -9
  636. transformers/models/hgnet_v2/modular_hgnet_v2.py +16 -9
  637. transformers/models/hiera/configuration_hiera.py +0 -1
  638. transformers/models/hiera/modeling_hiera.py +60 -62
  639. transformers/models/hubert/configuration_hubert.py +0 -1
  640. transformers/models/hubert/modeling_hubert.py +39 -37
  641. transformers/models/hubert/modular_hubert.py +12 -11
  642. transformers/models/hunyuan_v1_dense/configuration_hunyuan_v1_dense.py +21 -24
  643. transformers/models/hunyuan_v1_dense/modeling_hunyuan_v1_dense.py +31 -34
  644. transformers/models/hunyuan_v1_dense/modular_hunyuan_v1_dense.py +4 -6
  645. transformers/models/hunyuan_v1_moe/__init__.py +1 -1
  646. transformers/models/hunyuan_v1_moe/configuration_hunyuan_v1_moe.py +25 -28
  647. transformers/models/hunyuan_v1_moe/modeling_hunyuan_v1_moe.py +44 -39
  648. transformers/models/hunyuan_v1_moe/modular_hunyuan_v1_moe.py +9 -9
  649. transformers/models/ibert/configuration_ibert.py +0 -1
  650. transformers/models/ibert/modeling_ibert.py +76 -62
  651. transformers/models/ibert/quant_modules.py +0 -1
  652. transformers/models/idefics/configuration_idefics.py +0 -1
  653. transformers/models/idefics/image_processing_idefics.py +13 -15
  654. transformers/models/idefics/modeling_idefics.py +70 -61
  655. transformers/models/idefics/perceiver.py +1 -3
  656. transformers/models/idefics/processing_idefics.py +32 -48
  657. transformers/models/idefics/vision.py +22 -24
  658. transformers/models/idefics2/configuration_idefics2.py +0 -1
  659. transformers/models/idefics2/image_processing_idefics2.py +31 -32
  660. transformers/models/idefics2/image_processing_idefics2_fast.py +7 -8
  661. transformers/models/idefics2/modeling_idefics2.py +63 -59
  662. transformers/models/idefics2/processing_idefics2.py +10 -68
  663. transformers/models/idefics3/configuration_idefics3.py +0 -1
  664. transformers/models/idefics3/image_processing_idefics3.py +42 -43
  665. transformers/models/idefics3/image_processing_idefics3_fast.py +11 -12
  666. transformers/models/idefics3/modeling_idefics3.py +57 -55
  667. transformers/models/idefics3/processing_idefics3.py +15 -69
  668. transformers/models/ijepa/configuration_ijepa.py +0 -1
  669. transformers/models/ijepa/modeling_ijepa.py +10 -11
  670. transformers/models/ijepa/modular_ijepa.py +5 -7
  671. transformers/models/imagegpt/configuration_imagegpt.py +0 -1
  672. transformers/models/imagegpt/image_processing_imagegpt.py +17 -18
  673. transformers/models/imagegpt/image_processing_imagegpt_fast.py +9 -14
  674. transformers/models/imagegpt/modeling_imagegpt.py +66 -60
  675. transformers/models/informer/configuration_informer.py +6 -9
  676. transformers/models/informer/modeling_informer.py +84 -86
  677. transformers/models/informer/modular_informer.py +13 -16
  678. transformers/models/instructblip/configuration_instructblip.py +0 -1
  679. transformers/models/instructblip/modeling_instructblip.py +45 -44
  680. transformers/models/instructblip/processing_instructblip.py +10 -36
  681. transformers/models/instructblipvideo/configuration_instructblipvideo.py +0 -1
  682. transformers/models/instructblipvideo/modeling_instructblipvideo.py +107 -105
  683. transformers/models/instructblipvideo/modular_instructblipvideo.py +34 -36
  684. transformers/models/instructblipvideo/processing_instructblipvideo.py +14 -33
  685. transformers/models/instructblipvideo/video_processing_instructblipvideo.py +4 -6
  686. transformers/models/internvl/configuration_internvl.py +0 -1
  687. transformers/models/internvl/modeling_internvl.py +52 -51
  688. transformers/models/internvl/modular_internvl.py +24 -30
  689. transformers/models/internvl/processing_internvl.py +12 -45
  690. transformers/models/internvl/video_processing_internvl.py +8 -10
  691. transformers/models/jais2/__init__.py +27 -0
  692. transformers/models/jais2/configuration_jais2.py +150 -0
  693. transformers/models/jais2/modeling_jais2.py +484 -0
  694. transformers/models/jais2/modular_jais2.py +194 -0
  695. transformers/models/jamba/configuration_jamba.py +0 -1
  696. transformers/models/jamba/modeling_jamba.py +67 -65
  697. transformers/models/jamba/modular_jamba.py +54 -55
  698. transformers/models/janus/configuration_janus.py +0 -1
  699. transformers/models/janus/image_processing_janus.py +35 -37
  700. transformers/models/janus/image_processing_janus_fast.py +12 -14
  701. transformers/models/janus/modeling_janus.py +56 -50
  702. transformers/models/janus/modular_janus.py +76 -70
  703. transformers/models/janus/processing_janus.py +17 -43
  704. transformers/models/jetmoe/configuration_jetmoe.py +20 -23
  705. transformers/models/jetmoe/modeling_jetmoe.py +41 -44
  706. transformers/models/jetmoe/modular_jetmoe.py +31 -33
  707. transformers/models/kosmos2/configuration_kosmos2.py +0 -1
  708. transformers/models/kosmos2/modeling_kosmos2.py +159 -148
  709. transformers/models/kosmos2/processing_kosmos2.py +40 -55
  710. transformers/models/kosmos2_5/__init__.py +0 -1
  711. transformers/models/kosmos2_5/configuration_kosmos2_5.py +0 -1
  712. transformers/models/kosmos2_5/image_processing_kosmos2_5.py +10 -12
  713. transformers/models/kosmos2_5/image_processing_kosmos2_5_fast.py +4 -13
  714. transformers/models/kosmos2_5/modeling_kosmos2_5.py +118 -110
  715. transformers/models/kosmos2_5/processing_kosmos2_5.py +8 -29
  716. transformers/models/kyutai_speech_to_text/configuration_kyutai_speech_to_text.py +23 -25
  717. transformers/models/kyutai_speech_to_text/feature_extraction_kyutai_speech_to_text.py +12 -14
  718. transformers/models/kyutai_speech_to_text/modeling_kyutai_speech_to_text.py +67 -68
  719. transformers/models/kyutai_speech_to_text/modular_kyutai_speech_to_text.py +28 -22
  720. transformers/models/kyutai_speech_to_text/processing_kyutai_speech_to_text.py +2 -8
  721. transformers/models/lasr/configuration_lasr.py +5 -3
  722. transformers/models/lasr/feature_extraction_lasr.py +10 -12
  723. transformers/models/lasr/modeling_lasr.py +21 -23
  724. transformers/models/lasr/modular_lasr.py +16 -11
  725. transformers/models/lasr/processing_lasr.py +12 -8
  726. transformers/models/lasr/tokenization_lasr.py +2 -4
  727. transformers/models/layoutlm/configuration_layoutlm.py +0 -1
  728. transformers/models/layoutlm/modeling_layoutlm.py +72 -72
  729. transformers/models/layoutlmv2/configuration_layoutlmv2.py +0 -1
  730. transformers/models/layoutlmv2/image_processing_layoutlmv2.py +18 -21
  731. transformers/models/layoutlmv2/image_processing_layoutlmv2_fast.py +5 -7
  732. transformers/models/layoutlmv2/modeling_layoutlmv2.py +60 -50
  733. transformers/models/layoutlmv2/processing_layoutlmv2.py +14 -44
  734. transformers/models/layoutlmv2/tokenization_layoutlmv2.py +64 -74
  735. transformers/models/layoutlmv3/configuration_layoutlmv3.py +0 -1
  736. transformers/models/layoutlmv3/image_processing_layoutlmv3.py +24 -26
  737. transformers/models/layoutlmv3/image_processing_layoutlmv3_fast.py +7 -9
  738. transformers/models/layoutlmv3/modeling_layoutlmv3.py +78 -56
  739. transformers/models/layoutlmv3/processing_layoutlmv3.py +14 -46
  740. transformers/models/layoutlmv3/tokenization_layoutlmv3.py +64 -75
  741. transformers/models/layoutxlm/configuration_layoutxlm.py +0 -1
  742. transformers/models/layoutxlm/modular_layoutxlm.py +0 -1
  743. transformers/models/layoutxlm/processing_layoutxlm.py +14 -44
  744. transformers/models/layoutxlm/tokenization_layoutxlm.py +65 -76
  745. transformers/models/led/configuration_led.py +1 -4
  746. transformers/models/led/modeling_led.py +119 -267
  747. transformers/models/levit/configuration_levit.py +0 -1
  748. transformers/models/levit/image_processing_levit.py +19 -21
  749. transformers/models/levit/image_processing_levit_fast.py +0 -1
  750. transformers/models/levit/modeling_levit.py +35 -19
  751. transformers/models/lfm2/configuration_lfm2.py +22 -23
  752. transformers/models/lfm2/modeling_lfm2.py +43 -45
  753. transformers/models/lfm2/modular_lfm2.py +29 -29
  754. transformers/models/lfm2_moe/__init__.py +0 -1
  755. transformers/models/lfm2_moe/configuration_lfm2_moe.py +1 -2
  756. transformers/models/lfm2_moe/modeling_lfm2_moe.py +58 -49
  757. transformers/models/lfm2_moe/modular_lfm2_moe.py +13 -37
  758. transformers/models/lfm2_vl/configuration_lfm2_vl.py +4 -1
  759. transformers/models/lfm2_vl/image_processing_lfm2_vl_fast.py +34 -5
  760. transformers/models/lfm2_vl/modeling_lfm2_vl.py +42 -38
  761. transformers/models/lfm2_vl/modular_lfm2_vl.py +28 -29
  762. transformers/models/lfm2_vl/processing_lfm2_vl.py +96 -76
  763. transformers/models/lightglue/image_processing_lightglue.py +16 -15
  764. transformers/models/lightglue/image_processing_lightglue_fast.py +5 -6
  765. transformers/models/lightglue/modeling_lightglue.py +28 -30
  766. transformers/models/lightglue/modular_lightglue.py +28 -28
  767. transformers/models/lighton_ocr/__init__.py +28 -0
  768. transformers/models/lighton_ocr/configuration_lighton_ocr.py +128 -0
  769. transformers/models/lighton_ocr/modeling_lighton_ocr.py +460 -0
  770. transformers/models/lighton_ocr/modular_lighton_ocr.py +403 -0
  771. transformers/models/lighton_ocr/processing_lighton_ocr.py +229 -0
  772. transformers/models/lilt/configuration_lilt.py +0 -1
  773. transformers/models/lilt/modeling_lilt.py +72 -70
  774. transformers/models/llama/configuration_llama.py +21 -24
  775. transformers/models/llama/modeling_llama.py +32 -35
  776. transformers/models/llama/tokenization_llama.py +2 -4
  777. transformers/models/llama4/configuration_llama4.py +20 -22
  778. transformers/models/llama4/image_processing_llama4_fast.py +9 -11
  779. transformers/models/llama4/modeling_llama4.py +78 -75
  780. transformers/models/llama4/processing_llama4.py +33 -57
  781. transformers/models/llava/configuration_llava.py +0 -1
  782. transformers/models/llava/image_processing_llava.py +25 -28
  783. transformers/models/llava/image_processing_llava_fast.py +6 -8
  784. transformers/models/llava/modeling_llava.py +47 -44
  785. transformers/models/llava/processing_llava.py +18 -51
  786. transformers/models/llava_next/configuration_llava_next.py +0 -1
  787. transformers/models/llava_next/image_processing_llava_next.py +43 -45
  788. transformers/models/llava_next/image_processing_llava_next_fast.py +5 -7
  789. transformers/models/llava_next/modeling_llava_next.py +49 -47
  790. transformers/models/llava_next/processing_llava_next.py +18 -47
  791. transformers/models/llava_next_video/configuration_llava_next_video.py +0 -1
  792. transformers/models/llava_next_video/modeling_llava_next_video.py +60 -58
  793. transformers/models/llava_next_video/modular_llava_next_video.py +51 -49
  794. transformers/models/llava_next_video/processing_llava_next_video.py +21 -63
  795. transformers/models/llava_next_video/video_processing_llava_next_video.py +0 -1
  796. transformers/models/llava_onevision/configuration_llava_onevision.py +0 -1
  797. transformers/models/llava_onevision/image_processing_llava_onevision.py +40 -42
  798. transformers/models/llava_onevision/image_processing_llava_onevision_fast.py +6 -8
  799. transformers/models/llava_onevision/modeling_llava_onevision.py +67 -65
  800. transformers/models/llava_onevision/modular_llava_onevision.py +58 -56
  801. transformers/models/llava_onevision/processing_llava_onevision.py +21 -53
  802. transformers/models/llava_onevision/video_processing_llava_onevision.py +0 -1
  803. transformers/models/longcat_flash/__init__.py +0 -1
  804. transformers/models/longcat_flash/configuration_longcat_flash.py +32 -35
  805. transformers/models/longcat_flash/modeling_longcat_flash.py +32 -32
  806. transformers/models/longcat_flash/modular_longcat_flash.py +18 -19
  807. transformers/models/longformer/configuration_longformer.py +1 -4
  808. transformers/models/longformer/modeling_longformer.py +99 -101
  809. transformers/models/longt5/configuration_longt5.py +0 -1
  810. transformers/models/longt5/modeling_longt5.py +43 -48
  811. transformers/models/luke/configuration_luke.py +0 -1
  812. transformers/models/luke/modeling_luke.py +179 -181
  813. transformers/models/luke/tokenization_luke.py +99 -105
  814. transformers/models/lw_detr/__init__.py +27 -0
  815. transformers/models/lw_detr/configuration_lw_detr.py +374 -0
  816. transformers/models/lw_detr/modeling_lw_detr.py +1698 -0
  817. transformers/models/lw_detr/modular_lw_detr.py +1611 -0
  818. transformers/models/lxmert/configuration_lxmert.py +0 -1
  819. transformers/models/lxmert/modeling_lxmert.py +63 -74
  820. transformers/models/m2m_100/configuration_m2m_100.py +0 -1
  821. transformers/models/m2m_100/modeling_m2m_100.py +79 -71
  822. transformers/models/m2m_100/tokenization_m2m_100.py +8 -8
  823. transformers/models/mamba/configuration_mamba.py +0 -1
  824. transformers/models/mamba/modeling_mamba.py +44 -44
  825. transformers/models/mamba2/configuration_mamba2.py +0 -1
  826. transformers/models/mamba2/modeling_mamba2.py +67 -68
  827. transformers/models/marian/configuration_marian.py +1 -2
  828. transformers/models/marian/modeling_marian.py +87 -86
  829. transformers/models/marian/tokenization_marian.py +6 -6
  830. transformers/models/markuplm/configuration_markuplm.py +0 -1
  831. transformers/models/markuplm/feature_extraction_markuplm.py +1 -2
  832. transformers/models/markuplm/modeling_markuplm.py +65 -70
  833. transformers/models/markuplm/processing_markuplm.py +31 -38
  834. transformers/models/markuplm/tokenization_markuplm.py +67 -77
  835. transformers/models/mask2former/configuration_mask2former.py +5 -8
  836. transformers/models/mask2former/image_processing_mask2former.py +84 -85
  837. transformers/models/mask2former/image_processing_mask2former_fast.py +30 -33
  838. transformers/models/mask2former/modeling_mask2former.py +99 -92
  839. transformers/models/mask2former/modular_mask2former.py +6 -8
  840. transformers/models/maskformer/configuration_maskformer.py +6 -9
  841. transformers/models/maskformer/configuration_maskformer_swin.py +0 -1
  842. transformers/models/maskformer/image_processing_maskformer.py +84 -85
  843. transformers/models/maskformer/image_processing_maskformer_fast.py +29 -33
  844. transformers/models/maskformer/modeling_maskformer.py +65 -59
  845. transformers/models/maskformer/modeling_maskformer_swin.py +34 -32
  846. transformers/models/mbart/configuration_mbart.py +1 -1
  847. transformers/models/mbart/modeling_mbart.py +118 -113
  848. transformers/models/mbart/tokenization_mbart.py +2 -4
  849. transformers/models/mbart50/tokenization_mbart50.py +3 -5
  850. transformers/models/megatron_bert/configuration_megatron_bert.py +0 -1
  851. transformers/models/megatron_bert/modeling_megatron_bert.py +141 -150
  852. transformers/models/metaclip_2/modeling_metaclip_2.py +48 -46
  853. transformers/models/metaclip_2/modular_metaclip_2.py +21 -21
  854. transformers/models/mgp_str/configuration_mgp_str.py +0 -1
  855. transformers/models/mgp_str/modeling_mgp_str.py +14 -16
  856. transformers/models/mgp_str/processing_mgp_str.py +3 -20
  857. transformers/models/mgp_str/tokenization_mgp_str.py +1 -3
  858. transformers/models/mimi/configuration_mimi.py +38 -40
  859. transformers/models/mimi/modeling_mimi.py +100 -82
  860. transformers/models/minimax/__init__.py +0 -1
  861. transformers/models/minimax/configuration_minimax.py +32 -36
  862. transformers/models/minimax/modeling_minimax.py +57 -47
  863. transformers/models/minimax/modular_minimax.py +62 -54
  864. transformers/models/minimax_m2/__init__.py +28 -0
  865. transformers/models/minimax_m2/configuration_minimax_m2.py +211 -0
  866. transformers/models/minimax_m2/modeling_minimax_m2.py +704 -0
  867. transformers/models/minimax_m2/modular_minimax_m2.py +369 -0
  868. transformers/models/ministral/configuration_ministral.py +20 -22
  869. transformers/models/ministral/modeling_ministral.py +32 -34
  870. transformers/models/ministral/modular_ministral.py +27 -29
  871. transformers/models/ministral3/configuration_ministral3.py +19 -22
  872. transformers/models/ministral3/modeling_ministral3.py +32 -34
  873. transformers/models/ministral3/modular_ministral3.py +4 -5
  874. transformers/models/mistral/configuration_mistral.py +19 -22
  875. transformers/models/mistral/modeling_mistral.py +32 -34
  876. transformers/models/mistral/modular_mistral.py +11 -12
  877. transformers/models/mistral3/configuration_mistral3.py +0 -1
  878. transformers/models/mistral3/modeling_mistral3.py +53 -46
  879. transformers/models/mistral3/modular_mistral3.py +38 -36
  880. transformers/models/mixtral/configuration_mixtral.py +24 -27
  881. transformers/models/mixtral/modeling_mixtral.py +47 -42
  882. transformers/models/mixtral/modular_mixtral.py +32 -31
  883. transformers/models/mlcd/configuration_mlcd.py +0 -1
  884. transformers/models/mlcd/modeling_mlcd.py +16 -12
  885. transformers/models/mlcd/modular_mlcd.py +13 -11
  886. transformers/models/mllama/configuration_mllama.py +5 -8
  887. transformers/models/mllama/image_processing_mllama.py +23 -25
  888. transformers/models/mllama/image_processing_mllama_fast.py +5 -6
  889. transformers/models/mllama/modeling_mllama.py +94 -86
  890. transformers/models/mllama/processing_mllama.py +6 -55
  891. transformers/models/mluke/tokenization_mluke.py +97 -103
  892. transformers/models/mm_grounding_dino/configuration_mm_grounding_dino.py +1 -3
  893. transformers/models/mm_grounding_dino/modeling_mm_grounding_dino.py +95 -97
  894. transformers/models/mm_grounding_dino/modular_mm_grounding_dino.py +1 -3
  895. transformers/models/mobilebert/configuration_mobilebert.py +0 -1
  896. transformers/models/mobilebert/modeling_mobilebert.py +77 -85
  897. transformers/models/mobilebert/tokenization_mobilebert.py +0 -1
  898. transformers/models/mobilenet_v1/configuration_mobilenet_v1.py +0 -1
  899. transformers/models/mobilenet_v1/image_processing_mobilenet_v1.py +20 -23
  900. transformers/models/mobilenet_v1/image_processing_mobilenet_v1_fast.py +0 -1
  901. transformers/models/mobilenet_v1/modeling_mobilenet_v1.py +13 -16
  902. transformers/models/mobilenet_v2/configuration_mobilenet_v2.py +0 -1
  903. transformers/models/mobilenet_v2/image_processing_mobilenet_v2.py +48 -51
  904. transformers/models/mobilenet_v2/image_processing_mobilenet_v2_fast.py +10 -12
  905. transformers/models/mobilenet_v2/modeling_mobilenet_v2.py +17 -20
  906. transformers/models/mobilevit/configuration_mobilevit.py +0 -1
  907. transformers/models/mobilevit/image_processing_mobilevit.py +46 -49
  908. transformers/models/mobilevit/image_processing_mobilevit_fast.py +9 -11
  909. transformers/models/mobilevit/modeling_mobilevit.py +21 -19
  910. transformers/models/mobilevitv2/configuration_mobilevitv2.py +0 -1
  911. transformers/models/mobilevitv2/modeling_mobilevitv2.py +21 -20
  912. transformers/models/modernbert/configuration_modernbert.py +34 -34
  913. transformers/models/modernbert/modeling_modernbert.py +135 -126
  914. transformers/models/modernbert/modular_modernbert.py +167 -156
  915. transformers/models/modernbert_decoder/configuration_modernbert_decoder.py +30 -32
  916. transformers/models/modernbert_decoder/modeling_modernbert_decoder.py +54 -48
  917. transformers/models/modernbert_decoder/modular_modernbert_decoder.py +78 -71
  918. transformers/models/moonshine/configuration_moonshine.py +22 -24
  919. transformers/models/moonshine/modeling_moonshine.py +64 -66
  920. transformers/models/moonshine/modular_moonshine.py +72 -73
  921. transformers/models/moshi/configuration_moshi.py +18 -21
  922. transformers/models/moshi/modeling_moshi.py +150 -183
  923. transformers/models/mpnet/configuration_mpnet.py +0 -1
  924. transformers/models/mpnet/modeling_mpnet.py +57 -57
  925. transformers/models/mpnet/tokenization_mpnet.py +1 -4
  926. transformers/models/mpt/configuration_mpt.py +1 -9
  927. transformers/models/mpt/modeling_mpt.py +58 -60
  928. transformers/models/mra/configuration_mra.py +0 -1
  929. transformers/models/mra/modeling_mra.py +58 -57
  930. transformers/models/mt5/configuration_mt5.py +2 -4
  931. transformers/models/mt5/modeling_mt5.py +75 -87
  932. transformers/models/musicgen/configuration_musicgen.py +0 -1
  933. transformers/models/musicgen/modeling_musicgen.py +113 -120
  934. transformers/models/musicgen/processing_musicgen.py +3 -21
  935. transformers/models/musicgen_melody/configuration_musicgen_melody.py +0 -1
  936. transformers/models/musicgen_melody/feature_extraction_musicgen_melody.py +8 -9
  937. transformers/models/musicgen_melody/modeling_musicgen_melody.py +110 -109
  938. transformers/models/musicgen_melody/processing_musicgen_melody.py +3 -22
  939. transformers/models/mvp/configuration_mvp.py +0 -1
  940. transformers/models/mvp/modeling_mvp.py +122 -119
  941. transformers/models/myt5/tokenization_myt5.py +8 -10
  942. transformers/models/nanochat/configuration_nanochat.py +0 -1
  943. transformers/models/nanochat/modeling_nanochat.py +33 -36
  944. transformers/models/nanochat/modular_nanochat.py +12 -14
  945. transformers/models/nemotron/configuration_nemotron.py +20 -23
  946. transformers/models/nemotron/modeling_nemotron.py +51 -54
  947. transformers/models/nllb/tokenization_nllb.py +7 -9
  948. transformers/models/nllb_moe/configuration_nllb_moe.py +1 -1
  949. transformers/models/nllb_moe/modeling_nllb_moe.py +77 -69
  950. transformers/models/nougat/image_processing_nougat.py +29 -32
  951. transformers/models/nougat/image_processing_nougat_fast.py +4 -6
  952. transformers/models/nougat/processing_nougat.py +37 -39
  953. transformers/models/nougat/tokenization_nougat.py +16 -23
  954. transformers/models/nystromformer/configuration_nystromformer.py +0 -1
  955. transformers/models/nystromformer/modeling_nystromformer.py +68 -63
  956. transformers/models/olmo/configuration_olmo.py +18 -21
  957. transformers/models/olmo/modeling_olmo.py +32 -35
  958. transformers/models/olmo/modular_olmo.py +5 -9
  959. transformers/models/olmo2/configuration_olmo2.py +18 -21
  960. transformers/models/olmo2/modeling_olmo2.py +33 -36
  961. transformers/models/olmo2/modular_olmo2.py +29 -31
  962. transformers/models/olmo3/__init__.py +0 -1
  963. transformers/models/olmo3/configuration_olmo3.py +20 -23
  964. transformers/models/olmo3/modeling_olmo3.py +32 -35
  965. transformers/models/olmo3/modular_olmo3.py +31 -33
  966. transformers/models/olmoe/configuration_olmoe.py +24 -26
  967. transformers/models/olmoe/modeling_olmoe.py +49 -43
  968. transformers/models/olmoe/modular_olmoe.py +16 -15
  969. transformers/models/omdet_turbo/configuration_omdet_turbo.py +2 -3
  970. transformers/models/omdet_turbo/modeling_omdet_turbo.py +42 -40
  971. transformers/models/omdet_turbo/processing_omdet_turbo.py +19 -67
  972. transformers/models/oneformer/configuration_oneformer.py +5 -8
  973. transformers/models/oneformer/image_processing_oneformer.py +83 -84
  974. transformers/models/oneformer/image_processing_oneformer_fast.py +33 -34
  975. transformers/models/oneformer/modeling_oneformer.py +130 -162
  976. transformers/models/oneformer/processing_oneformer.py +28 -43
  977. transformers/models/openai/configuration_openai.py +0 -1
  978. transformers/models/openai/modeling_openai.py +62 -51
  979. transformers/models/openai/tokenization_openai.py +2 -5
  980. transformers/models/opt/configuration_opt.py +0 -1
  981. transformers/models/opt/modeling_opt.py +74 -75
  982. transformers/models/ovis2/__init__.py +0 -1
  983. transformers/models/ovis2/configuration_ovis2.py +0 -1
  984. transformers/models/ovis2/image_processing_ovis2.py +22 -24
  985. transformers/models/ovis2/image_processing_ovis2_fast.py +6 -8
  986. transformers/models/ovis2/modeling_ovis2.py +58 -48
  987. transformers/models/ovis2/modular_ovis2.py +38 -32
  988. transformers/models/ovis2/processing_ovis2.py +12 -40
  989. transformers/models/owlv2/configuration_owlv2.py +0 -1
  990. transformers/models/owlv2/image_processing_owlv2.py +20 -21
  991. transformers/models/owlv2/image_processing_owlv2_fast.py +7 -10
  992. transformers/models/owlv2/modeling_owlv2.py +89 -90
  993. transformers/models/owlv2/modular_owlv2.py +6 -9
  994. transformers/models/owlv2/processing_owlv2.py +20 -49
  995. transformers/models/owlvit/configuration_owlvit.py +0 -1
  996. transformers/models/owlvit/image_processing_owlvit.py +21 -22
  997. transformers/models/owlvit/image_processing_owlvit_fast.py +2 -3
  998. transformers/models/owlvit/modeling_owlvit.py +88 -89
  999. transformers/models/owlvit/processing_owlvit.py +20 -48
  1000. transformers/models/paddleocr_vl/__init__.py +0 -1
  1001. transformers/models/paddleocr_vl/configuration_paddleocr_vl.py +19 -19
  1002. transformers/models/paddleocr_vl/image_processing_paddleocr_vl.py +37 -37
  1003. transformers/models/paddleocr_vl/image_processing_paddleocr_vl_fast.py +12 -12
  1004. transformers/models/paddleocr_vl/modeling_paddleocr_vl.py +104 -90
  1005. transformers/models/paddleocr_vl/modular_paddleocr_vl.py +90 -80
  1006. transformers/models/paddleocr_vl/processing_paddleocr_vl.py +1 -3
  1007. transformers/models/paligemma/configuration_paligemma.py +0 -1
  1008. transformers/models/paligemma/modeling_paligemma.py +73 -67
  1009. transformers/models/paligemma/processing_paligemma.py +13 -66
  1010. transformers/models/parakeet/configuration_parakeet.py +1 -4
  1011. transformers/models/parakeet/feature_extraction_parakeet.py +10 -12
  1012. transformers/models/parakeet/modeling_parakeet.py +23 -22
  1013. transformers/models/parakeet/modular_parakeet.py +21 -18
  1014. transformers/models/parakeet/processing_parakeet.py +12 -5
  1015. transformers/models/parakeet/{tokenization_parakeet_fast.py → tokenization_parakeet.py} +5 -7
  1016. transformers/models/patchtsmixer/configuration_patchtsmixer.py +5 -8
  1017. transformers/models/patchtsmixer/modeling_patchtsmixer.py +64 -62
  1018. transformers/models/patchtst/configuration_patchtst.py +6 -9
  1019. transformers/models/patchtst/modeling_patchtst.py +77 -78
  1020. transformers/models/pe_audio/__init__.py +29 -0
  1021. transformers/models/pe_audio/configuration_pe_audio.py +204 -0
  1022. transformers/models/pe_audio/feature_extraction_pe_audio.py +160 -0
  1023. transformers/models/pe_audio/modeling_pe_audio.py +819 -0
  1024. transformers/models/pe_audio/modular_pe_audio.py +298 -0
  1025. transformers/models/pe_audio/processing_pe_audio.py +23 -0
  1026. transformers/models/pe_audio_video/__init__.py +28 -0
  1027. transformers/models/pe_audio_video/configuration_pe_audio_video.py +223 -0
  1028. transformers/models/pe_audio_video/modeling_pe_audio_video.py +971 -0
  1029. transformers/models/pe_audio_video/modular_pe_audio_video.py +763 -0
  1030. transformers/models/pe_audio_video/processing_pe_audio_video.py +24 -0
  1031. transformers/models/pe_video/__init__.py +29 -0
  1032. transformers/models/pe_video/configuration_pe_video.py +209 -0
  1033. transformers/models/pe_video/modeling_pe_video.py +635 -0
  1034. transformers/models/pe_video/modular_pe_video.py +218 -0
  1035. transformers/models/pe_video/processing_pe_video.py +10 -0
  1036. transformers/models/pe_video/video_processing_pe_video.py +64 -0
  1037. transformers/models/pegasus/configuration_pegasus.py +1 -1
  1038. transformers/models/pegasus/modeling_pegasus.py +66 -65
  1039. transformers/models/pegasus/tokenization_pegasus.py +1 -4
  1040. transformers/models/pegasus_x/configuration_pegasus_x.py +0 -1
  1041. transformers/models/pegasus_x/modeling_pegasus_x.py +51 -52
  1042. transformers/models/perceiver/configuration_perceiver.py +0 -1
  1043. transformers/models/perceiver/image_processing_perceiver.py +22 -25
  1044. transformers/models/perceiver/image_processing_perceiver_fast.py +5 -7
  1045. transformers/models/perceiver/modeling_perceiver.py +140 -137
  1046. transformers/models/perceiver/tokenization_perceiver.py +3 -6
  1047. transformers/models/perception_lm/configuration_perception_lm.py +0 -1
  1048. transformers/models/perception_lm/image_processing_perception_lm_fast.py +8 -10
  1049. transformers/models/perception_lm/modeling_perception_lm.py +45 -43
  1050. transformers/models/perception_lm/modular_perception_lm.py +38 -36
  1051. transformers/models/perception_lm/processing_perception_lm.py +13 -47
  1052. transformers/models/perception_lm/video_processing_perception_lm.py +0 -1
  1053. transformers/models/persimmon/configuration_persimmon.py +18 -21
  1054. transformers/models/persimmon/modeling_persimmon.py +40 -43
  1055. transformers/models/phi/configuration_phi.py +19 -22
  1056. transformers/models/phi/modeling_phi.py +36 -38
  1057. transformers/models/phi/modular_phi.py +23 -23
  1058. transformers/models/phi3/configuration_phi3.py +23 -26
  1059. transformers/models/phi3/modeling_phi3.py +34 -37
  1060. transformers/models/phi3/modular_phi3.py +13 -17
  1061. transformers/models/phi4_multimodal/configuration_phi4_multimodal.py +25 -26
  1062. transformers/models/phi4_multimodal/feature_extraction_phi4_multimodal.py +7 -9
  1063. transformers/models/phi4_multimodal/image_processing_phi4_multimodal_fast.py +7 -7
  1064. transformers/models/phi4_multimodal/modeling_phi4_multimodal.py +58 -57
  1065. transformers/models/phi4_multimodal/modular_phi4_multimodal.py +62 -60
  1066. transformers/models/phi4_multimodal/processing_phi4_multimodal.py +7 -44
  1067. transformers/models/phimoe/configuration_phimoe.py +26 -29
  1068. transformers/models/phimoe/modeling_phimoe.py +47 -42
  1069. transformers/models/phimoe/modular_phimoe.py +1 -2
  1070. transformers/models/phobert/tokenization_phobert.py +4 -6
  1071. transformers/models/pix2struct/configuration_pix2struct.py +0 -1
  1072. transformers/models/pix2struct/image_processing_pix2struct.py +15 -19
  1073. transformers/models/pix2struct/image_processing_pix2struct_fast.py +7 -10
  1074. transformers/models/pix2struct/modeling_pix2struct.py +42 -45
  1075. transformers/models/pix2struct/processing_pix2struct.py +5 -30
  1076. transformers/models/pixio/__init__.py +29 -0
  1077. transformers/models/pixio/configuration_pixio.py +150 -0
  1078. transformers/models/pixio/modeling_pixio.py +505 -0
  1079. transformers/models/pixio/modular_pixio.py +401 -0
  1080. transformers/models/pixtral/configuration_pixtral.py +11 -14
  1081. transformers/models/pixtral/image_processing_pixtral.py +26 -28
  1082. transformers/models/pixtral/image_processing_pixtral_fast.py +5 -6
  1083. transformers/models/pixtral/modeling_pixtral.py +23 -26
  1084. transformers/models/pixtral/processing_pixtral.py +21 -53
  1085. transformers/models/plbart/configuration_plbart.py +1 -1
  1086. transformers/models/plbart/modeling_plbart.py +107 -102
  1087. transformers/models/plbart/modular_plbart.py +36 -32
  1088. transformers/models/plbart/tokenization_plbart.py +4 -5
  1089. transformers/models/poolformer/configuration_poolformer.py +0 -1
  1090. transformers/models/poolformer/image_processing_poolformer.py +21 -24
  1091. transformers/models/poolformer/image_processing_poolformer_fast.py +6 -8
  1092. transformers/models/poolformer/modeling_poolformer.py +21 -13
  1093. transformers/models/pop2piano/configuration_pop2piano.py +0 -2
  1094. transformers/models/pop2piano/feature_extraction_pop2piano.py +6 -9
  1095. transformers/models/pop2piano/modeling_pop2piano.py +22 -23
  1096. transformers/models/pop2piano/processing_pop2piano.py +25 -33
  1097. transformers/models/pop2piano/tokenization_pop2piano.py +15 -23
  1098. transformers/models/prompt_depth_anything/configuration_prompt_depth_anything.py +3 -3
  1099. transformers/models/prompt_depth_anything/image_processing_prompt_depth_anything.py +28 -28
  1100. transformers/models/prompt_depth_anything/image_processing_prompt_depth_anything_fast.py +14 -15
  1101. transformers/models/prompt_depth_anything/modeling_prompt_depth_anything.py +9 -10
  1102. transformers/models/prompt_depth_anything/modular_prompt_depth_anything.py +9 -10
  1103. transformers/models/prophetnet/configuration_prophetnet.py +26 -28
  1104. transformers/models/prophetnet/modeling_prophetnet.py +111 -131
  1105. transformers/models/prophetnet/tokenization_prophetnet.py +14 -16
  1106. transformers/models/pvt/configuration_pvt.py +0 -1
  1107. transformers/models/pvt/image_processing_pvt.py +17 -20
  1108. transformers/models/pvt/image_processing_pvt_fast.py +0 -1
  1109. transformers/models/pvt/modeling_pvt.py +19 -21
  1110. transformers/models/pvt_v2/configuration_pvt_v2.py +2 -4
  1111. transformers/models/pvt_v2/modeling_pvt_v2.py +21 -23
  1112. transformers/models/qwen2/configuration_qwen2.py +18 -21
  1113. transformers/models/qwen2/modeling_qwen2.py +32 -34
  1114. transformers/models/qwen2/modular_qwen2.py +11 -12
  1115. transformers/models/qwen2/tokenization_qwen2.py +2 -5
  1116. transformers/models/qwen2_5_omni/configuration_qwen2_5_omni.py +20 -23
  1117. transformers/models/qwen2_5_omni/modeling_qwen2_5_omni.py +239 -192
  1118. transformers/models/qwen2_5_omni/modular_qwen2_5_omni.py +174 -127
  1119. transformers/models/qwen2_5_omni/processing_qwen2_5_omni.py +41 -49
  1120. transformers/models/qwen2_5_vl/configuration_qwen2_5_vl.py +22 -25
  1121. transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py +112 -101
  1122. transformers/models/qwen2_5_vl/modular_qwen2_5_vl.py +72 -107
  1123. transformers/models/qwen2_5_vl/processing_qwen2_5_vl.py +7 -43
  1124. transformers/models/qwen2_audio/configuration_qwen2_audio.py +0 -1
  1125. transformers/models/qwen2_audio/modeling_qwen2_audio.py +29 -31
  1126. transformers/models/qwen2_audio/processing_qwen2_audio.py +13 -42
  1127. transformers/models/qwen2_moe/configuration_qwen2_moe.py +28 -31
  1128. transformers/models/qwen2_moe/modeling_qwen2_moe.py +48 -43
  1129. transformers/models/qwen2_moe/modular_qwen2_moe.py +7 -10
  1130. transformers/models/qwen2_vl/configuration_qwen2_vl.py +22 -24
  1131. transformers/models/qwen2_vl/image_processing_qwen2_vl.py +41 -42
  1132. transformers/models/qwen2_vl/image_processing_qwen2_vl_fast.py +8 -9
  1133. transformers/models/qwen2_vl/modeling_qwen2_vl.py +108 -96
  1134. transformers/models/qwen2_vl/processing_qwen2_vl.py +7 -44
  1135. transformers/models/qwen2_vl/video_processing_qwen2_vl.py +35 -13
  1136. transformers/models/qwen3/configuration_qwen3.py +20 -23
  1137. transformers/models/qwen3/modeling_qwen3.py +32 -35
  1138. transformers/models/qwen3/modular_qwen3.py +4 -6
  1139. transformers/models/qwen3_moe/configuration_qwen3_moe.py +25 -28
  1140. transformers/models/qwen3_moe/modeling_qwen3_moe.py +48 -43
  1141. transformers/models/qwen3_moe/modular_qwen3_moe.py +10 -13
  1142. transformers/models/qwen3_next/configuration_qwen3_next.py +31 -34
  1143. transformers/models/qwen3_next/modeling_qwen3_next.py +43 -48
  1144. transformers/models/qwen3_next/modular_qwen3_next.py +33 -34
  1145. transformers/models/qwen3_omni_moe/configuration_qwen3_omni_moe.py +89 -88
  1146. transformers/models/qwen3_omni_moe/modeling_qwen3_omni_moe.py +199 -156
  1147. transformers/models/qwen3_omni_moe/modular_qwen3_omni_moe.py +170 -152
  1148. transformers/models/qwen3_omni_moe/processing_qwen3_omni_moe.py +40 -48
  1149. transformers/models/qwen3_vl/configuration_qwen3_vl.py +21 -24
  1150. transformers/models/qwen3_vl/modeling_qwen3_vl.py +91 -81
  1151. transformers/models/qwen3_vl/modular_qwen3_vl.py +86 -112
  1152. transformers/models/qwen3_vl/processing_qwen3_vl.py +6 -42
  1153. transformers/models/qwen3_vl/video_processing_qwen3_vl.py +10 -12
  1154. transformers/models/qwen3_vl_moe/configuration_qwen3_vl_moe.py +21 -25
  1155. transformers/models/qwen3_vl_moe/modeling_qwen3_vl_moe.py +174 -195
  1156. transformers/models/qwen3_vl_moe/modular_qwen3_vl_moe.py +65 -117
  1157. transformers/models/rag/configuration_rag.py +0 -9
  1158. transformers/models/rag/modeling_rag.py +123 -127
  1159. transformers/models/rag/retrieval_rag.py +2 -4
  1160. transformers/models/rag/tokenization_rag.py +0 -50
  1161. transformers/models/recurrent_gemma/configuration_recurrent_gemma.py +21 -24
  1162. transformers/models/recurrent_gemma/modeling_recurrent_gemma.py +34 -36
  1163. transformers/models/reformer/configuration_reformer.py +0 -1
  1164. transformers/models/reformer/modeling_reformer.py +76 -69
  1165. transformers/models/reformer/tokenization_reformer.py +3 -6
  1166. transformers/models/regnet/configuration_regnet.py +0 -1
  1167. transformers/models/regnet/modeling_regnet.py +11 -9
  1168. transformers/models/rembert/configuration_rembert.py +0 -1
  1169. transformers/models/rembert/modeling_rembert.py +115 -111
  1170. transformers/models/rembert/tokenization_rembert.py +1 -4
  1171. transformers/models/resnet/configuration_resnet.py +0 -1
  1172. transformers/models/resnet/modeling_resnet.py +16 -13
  1173. transformers/models/roberta/configuration_roberta.py +0 -1
  1174. transformers/models/roberta/modeling_roberta.py +94 -93
  1175. transformers/models/roberta/modular_roberta.py +58 -58
  1176. transformers/models/roberta/tokenization_roberta.py +2 -5
  1177. transformers/models/roberta/tokenization_roberta_old.py +2 -4
  1178. transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py +0 -1
  1179. transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py +94 -93
  1180. transformers/models/roc_bert/configuration_roc_bert.py +0 -1
  1181. transformers/models/roc_bert/modeling_roc_bert.py +122 -121
  1182. transformers/models/roc_bert/tokenization_roc_bert.py +88 -94
  1183. transformers/models/roformer/configuration_roformer.py +0 -1
  1184. transformers/models/roformer/modeling_roformer.py +79 -81
  1185. transformers/models/roformer/tokenization_roformer.py +3 -6
  1186. transformers/models/roformer/tokenization_utils.py +0 -1
  1187. transformers/models/rt_detr/configuration_rt_detr.py +1 -2
  1188. transformers/models/rt_detr/configuration_rt_detr_resnet.py +0 -1
  1189. transformers/models/rt_detr/image_processing_rt_detr.py +54 -55
  1190. transformers/models/rt_detr/image_processing_rt_detr_fast.py +15 -15
  1191. transformers/models/rt_detr/modeling_rt_detr.py +84 -82
  1192. transformers/models/rt_detr/modeling_rt_detr_resnet.py +10 -7
  1193. transformers/models/rt_detr/modular_rt_detr.py +14 -14
  1194. transformers/models/rt_detr_v2/configuration_rt_detr_v2.py +2 -4
  1195. transformers/models/rt_detr_v2/modeling_rt_detr_v2.py +86 -81
  1196. transformers/models/rt_detr_v2/modular_rt_detr_v2.py +10 -7
  1197. transformers/models/rwkv/configuration_rwkv.py +0 -1
  1198. transformers/models/rwkv/modeling_rwkv.py +30 -32
  1199. transformers/models/sam/configuration_sam.py +1 -1
  1200. transformers/models/sam/image_processing_sam.py +59 -60
  1201. transformers/models/sam/image_processing_sam_fast.py +21 -23
  1202. transformers/models/sam/modeling_sam.py +37 -36
  1203. transformers/models/sam/processing_sam.py +39 -27
  1204. transformers/models/sam2/configuration_sam2.py +1 -2
  1205. transformers/models/sam2/image_processing_sam2_fast.py +14 -15
  1206. transformers/models/sam2/modeling_sam2.py +50 -48
  1207. transformers/models/sam2/modular_sam2.py +48 -45
  1208. transformers/models/sam2/processing_sam2.py +31 -47
  1209. transformers/models/sam2_video/configuration_sam2_video.py +0 -1
  1210. transformers/models/sam2_video/modeling_sam2_video.py +119 -112
  1211. transformers/models/sam2_video/modular_sam2_video.py +91 -97
  1212. transformers/models/sam2_video/processing_sam2_video.py +49 -66
  1213. transformers/models/sam2_video/video_processing_sam2_video.py +1 -4
  1214. transformers/models/sam3/configuration_sam3.py +21 -2
  1215. transformers/models/sam3/image_processing_sam3_fast.py +17 -20
  1216. transformers/models/sam3/modeling_sam3.py +77 -56
  1217. transformers/models/sam3/modular_sam3.py +3 -8
  1218. transformers/models/sam3/processing_sam3.py +29 -48
  1219. transformers/models/sam3_tracker/__init__.py +0 -1
  1220. transformers/models/sam3_tracker/configuration_sam3_tracker.py +0 -1
  1221. transformers/models/sam3_tracker/modeling_sam3_tracker.py +36 -36
  1222. transformers/models/sam3_tracker/modular_sam3_tracker.py +2 -1
  1223. transformers/models/sam3_tracker/processing_sam3_tracker.py +31 -47
  1224. transformers/models/sam3_tracker_video/__init__.py +0 -1
  1225. transformers/models/sam3_tracker_video/configuration_sam3_tracker_video.py +25 -1
  1226. transformers/models/sam3_tracker_video/modeling_sam3_tracker_video.py +96 -85
  1227. transformers/models/sam3_tracker_video/modular_sam3_tracker_video.py +27 -6
  1228. transformers/models/sam3_tracker_video/processing_sam3_tracker_video.py +50 -66
  1229. transformers/models/sam3_video/configuration_sam3_video.py +14 -1
  1230. transformers/models/sam3_video/modeling_sam3_video.py +32 -34
  1231. transformers/models/sam3_video/processing_sam3_video.py +26 -46
  1232. transformers/models/sam_hq/__init__.py +1 -1
  1233. transformers/models/sam_hq/configuration_sam_hq.py +1 -1
  1234. transformers/models/sam_hq/modeling_sam_hq.py +65 -64
  1235. transformers/models/sam_hq/modular_sam_hq.py +17 -19
  1236. transformers/models/sam_hq/{processing_samhq.py → processing_sam_hq.py} +39 -28
  1237. transformers/models/seamless_m4t/configuration_seamless_m4t.py +0 -1
  1238. transformers/models/seamless_m4t/feature_extraction_seamless_m4t.py +8 -11
  1239. transformers/models/seamless_m4t/modeling_seamless_m4t.py +207 -193
  1240. transformers/models/seamless_m4t/processing_seamless_m4t.py +18 -39
  1241. transformers/models/seamless_m4t/tokenization_seamless_m4t.py +15 -20
  1242. transformers/models/seamless_m4t_v2/configuration_seamless_m4t_v2.py +0 -1
  1243. transformers/models/seamless_m4t_v2/modeling_seamless_m4t_v2.py +199 -195
  1244. transformers/models/seed_oss/configuration_seed_oss.py +23 -25
  1245. transformers/models/seed_oss/modeling_seed_oss.py +31 -33
  1246. transformers/models/seed_oss/modular_seed_oss.py +3 -4
  1247. transformers/models/segformer/configuration_segformer.py +0 -10
  1248. transformers/models/segformer/image_processing_segformer.py +39 -42
  1249. transformers/models/segformer/image_processing_segformer_fast.py +7 -9
  1250. transformers/models/segformer/modeling_segformer.py +26 -28
  1251. transformers/models/segformer/modular_segformer.py +5 -7
  1252. transformers/models/seggpt/configuration_seggpt.py +0 -1
  1253. transformers/models/seggpt/image_processing_seggpt.py +38 -41
  1254. transformers/models/seggpt/modeling_seggpt.py +28 -30
  1255. transformers/models/sew/configuration_sew.py +0 -1
  1256. transformers/models/sew/modeling_sew.py +33 -35
  1257. transformers/models/sew/modular_sew.py +10 -12
  1258. transformers/models/sew_d/configuration_sew_d.py +0 -1
  1259. transformers/models/sew_d/modeling_sew_d.py +28 -30
  1260. transformers/models/shieldgemma2/configuration_shieldgemma2.py +0 -1
  1261. transformers/models/shieldgemma2/modeling_shieldgemma2.py +16 -17
  1262. transformers/models/shieldgemma2/processing_shieldgemma2.py +3 -5
  1263. transformers/models/siglip/configuration_siglip.py +0 -1
  1264. transformers/models/siglip/image_processing_siglip.py +17 -20
  1265. transformers/models/siglip/image_processing_siglip_fast.py +0 -1
  1266. transformers/models/siglip/modeling_siglip.py +62 -41
  1267. transformers/models/siglip/processing_siglip.py +2 -14
  1268. transformers/models/siglip/tokenization_siglip.py +6 -7
  1269. transformers/models/siglip2/configuration_siglip2.py +1 -1
  1270. transformers/models/siglip2/image_processing_siglip2.py +15 -16
  1271. transformers/models/siglip2/image_processing_siglip2_fast.py +4 -5
  1272. transformers/models/siglip2/modeling_siglip2.py +114 -92
  1273. transformers/models/siglip2/modular_siglip2.py +23 -25
  1274. transformers/models/siglip2/processing_siglip2.py +2 -14
  1275. transformers/models/smollm3/configuration_smollm3.py +23 -26
  1276. transformers/models/smollm3/modeling_smollm3.py +32 -35
  1277. transformers/models/smollm3/modular_smollm3.py +27 -29
  1278. transformers/models/smolvlm/configuration_smolvlm.py +1 -1
  1279. transformers/models/smolvlm/image_processing_smolvlm.py +42 -43
  1280. transformers/models/smolvlm/image_processing_smolvlm_fast.py +12 -12
  1281. transformers/models/smolvlm/modeling_smolvlm.py +56 -53
  1282. transformers/models/smolvlm/modular_smolvlm.py +15 -17
  1283. transformers/models/smolvlm/processing_smolvlm.py +15 -76
  1284. transformers/models/smolvlm/video_processing_smolvlm.py +7 -9
  1285. transformers/models/speech_encoder_decoder/configuration_speech_encoder_decoder.py +0 -1
  1286. transformers/models/speech_encoder_decoder/modeling_speech_encoder_decoder.py +20 -23
  1287. transformers/models/speech_to_text/configuration_speech_to_text.py +0 -1
  1288. transformers/models/speech_to_text/feature_extraction_speech_to_text.py +10 -13
  1289. transformers/models/speech_to_text/modeling_speech_to_text.py +62 -54
  1290. transformers/models/speech_to_text/processing_speech_to_text.py +4 -30
  1291. transformers/models/speech_to_text/tokenization_speech_to_text.py +5 -6
  1292. transformers/models/speecht5/configuration_speecht5.py +0 -1
  1293. transformers/models/speecht5/feature_extraction_speecht5.py +16 -37
  1294. transformers/models/speecht5/modeling_speecht5.py +200 -174
  1295. transformers/models/speecht5/number_normalizer.py +0 -1
  1296. transformers/models/speecht5/processing_speecht5.py +3 -37
  1297. transformers/models/speecht5/tokenization_speecht5.py +4 -5
  1298. transformers/models/splinter/configuration_splinter.py +0 -1
  1299. transformers/models/splinter/modeling_splinter.py +63 -59
  1300. transformers/models/splinter/tokenization_splinter.py +2 -4
  1301. transformers/models/squeezebert/configuration_squeezebert.py +0 -1
  1302. transformers/models/squeezebert/modeling_squeezebert.py +62 -62
  1303. transformers/models/squeezebert/tokenization_squeezebert.py +0 -1
  1304. transformers/models/stablelm/configuration_stablelm.py +20 -23
  1305. transformers/models/stablelm/modeling_stablelm.py +40 -43
  1306. transformers/models/starcoder2/configuration_starcoder2.py +19 -22
  1307. transformers/models/starcoder2/modeling_starcoder2.py +34 -37
  1308. transformers/models/starcoder2/modular_starcoder2.py +13 -15
  1309. transformers/models/superglue/configuration_superglue.py +3 -3
  1310. transformers/models/superglue/image_processing_superglue.py +15 -15
  1311. transformers/models/superglue/image_processing_superglue_fast.py +5 -7
  1312. transformers/models/superglue/modeling_superglue.py +32 -33
  1313. transformers/models/superpoint/image_processing_superpoint.py +15 -15
  1314. transformers/models/superpoint/image_processing_superpoint_fast.py +5 -7
  1315. transformers/models/superpoint/modeling_superpoint.py +13 -14
  1316. transformers/models/swiftformer/configuration_swiftformer.py +0 -1
  1317. transformers/models/swiftformer/modeling_swiftformer.py +16 -14
  1318. transformers/models/swin/configuration_swin.py +0 -1
  1319. transformers/models/swin/modeling_swin.py +74 -82
  1320. transformers/models/swin2sr/configuration_swin2sr.py +0 -1
  1321. transformers/models/swin2sr/image_processing_swin2sr.py +10 -13
  1322. transformers/models/swin2sr/image_processing_swin2sr_fast.py +2 -6
  1323. transformers/models/swin2sr/modeling_swin2sr.py +75 -61
  1324. transformers/models/swinv2/configuration_swinv2.py +0 -1
  1325. transformers/models/swinv2/modeling_swinv2.py +96 -100
  1326. transformers/models/switch_transformers/configuration_switch_transformers.py +0 -1
  1327. transformers/models/switch_transformers/modeling_switch_transformers.py +34 -41
  1328. transformers/models/switch_transformers/modular_switch_transformers.py +31 -38
  1329. transformers/models/t5/configuration_t5.py +7 -2
  1330. transformers/models/t5/modeling_t5.py +76 -84
  1331. transformers/models/t5/tokenization_t5.py +1 -3
  1332. transformers/models/t5gemma/configuration_t5gemma.py +33 -34
  1333. transformers/models/t5gemma/modeling_t5gemma.py +97 -100
  1334. transformers/models/t5gemma/modular_t5gemma.py +117 -118
  1335. transformers/models/t5gemma2/configuration_t5gemma2.py +59 -96
  1336. transformers/models/t5gemma2/modeling_t5gemma2.py +109 -103
  1337. transformers/models/t5gemma2/modular_t5gemma2.py +375 -91
  1338. transformers/models/table_transformer/configuration_table_transformer.py +1 -2
  1339. transformers/models/table_transformer/modeling_table_transformer.py +47 -49
  1340. transformers/models/tapas/configuration_tapas.py +0 -1
  1341. transformers/models/tapas/modeling_tapas.py +64 -66
  1342. transformers/models/tapas/tokenization_tapas.py +115 -153
  1343. transformers/models/textnet/configuration_textnet.py +0 -1
  1344. transformers/models/textnet/image_processing_textnet.py +22 -25
  1345. transformers/models/textnet/image_processing_textnet_fast.py +5 -7
  1346. transformers/models/textnet/modeling_textnet.py +13 -14
  1347. transformers/models/time_series_transformer/configuration_time_series_transformer.py +5 -8
  1348. transformers/models/time_series_transformer/modeling_time_series_transformer.py +79 -81
  1349. transformers/models/timesfm/configuration_timesfm.py +0 -1
  1350. transformers/models/timesfm/modeling_timesfm.py +29 -19
  1351. transformers/models/timesfm/modular_timesfm.py +28 -18
  1352. transformers/models/timesformer/configuration_timesformer.py +0 -1
  1353. transformers/models/timesformer/modeling_timesformer.py +13 -16
  1354. transformers/models/timm_backbone/configuration_timm_backbone.py +0 -1
  1355. transformers/models/timm_backbone/modeling_timm_backbone.py +17 -15
  1356. transformers/models/timm_wrapper/configuration_timm_wrapper.py +5 -3
  1357. transformers/models/timm_wrapper/image_processing_timm_wrapper.py +4 -5
  1358. transformers/models/timm_wrapper/modeling_timm_wrapper.py +32 -28
  1359. transformers/models/trocr/configuration_trocr.py +0 -1
  1360. transformers/models/trocr/modeling_trocr.py +39 -42
  1361. transformers/models/trocr/processing_trocr.py +5 -25
  1362. transformers/models/tvp/configuration_tvp.py +5 -2
  1363. transformers/models/tvp/image_processing_tvp.py +50 -52
  1364. transformers/models/tvp/image_processing_tvp_fast.py +9 -10
  1365. transformers/models/tvp/modeling_tvp.py +25 -27
  1366. transformers/models/tvp/processing_tvp.py +2 -14
  1367. transformers/models/udop/configuration_udop.py +1 -1
  1368. transformers/models/udop/modeling_udop.py +63 -70
  1369. transformers/models/udop/processing_udop.py +7 -26
  1370. transformers/models/udop/tokenization_udop.py +80 -93
  1371. transformers/models/umt5/configuration_umt5.py +2 -3
  1372. transformers/models/umt5/modeling_umt5.py +80 -87
  1373. transformers/models/unispeech/configuration_unispeech.py +0 -1
  1374. transformers/models/unispeech/modeling_unispeech.py +47 -49
  1375. transformers/models/unispeech/modular_unispeech.py +20 -22
  1376. transformers/models/unispeech_sat/configuration_unispeech_sat.py +0 -1
  1377. transformers/models/unispeech_sat/modeling_unispeech_sat.py +63 -65
  1378. transformers/models/unispeech_sat/modular_unispeech_sat.py +21 -23
  1379. transformers/models/univnet/feature_extraction_univnet.py +14 -14
  1380. transformers/models/univnet/modeling_univnet.py +7 -8
  1381. transformers/models/upernet/configuration_upernet.py +0 -1
  1382. transformers/models/upernet/modeling_upernet.py +10 -13
  1383. transformers/models/vaultgemma/__init__.py +0 -1
  1384. transformers/models/vaultgemma/configuration_vaultgemma.py +24 -26
  1385. transformers/models/vaultgemma/modeling_vaultgemma.py +35 -37
  1386. transformers/models/vaultgemma/modular_vaultgemma.py +29 -31
  1387. transformers/models/video_llama_3/image_processing_video_llama_3.py +43 -42
  1388. transformers/models/video_llama_3/image_processing_video_llama_3_fast.py +8 -8
  1389. transformers/models/video_llama_3/modeling_video_llama_3.py +77 -66
  1390. transformers/models/video_llama_3/modular_video_llama_3.py +110 -112
  1391. transformers/models/video_llama_3/processing_video_llama_3.py +5 -39
  1392. transformers/models/video_llama_3/video_processing_video_llama_3.py +18 -18
  1393. transformers/models/video_llava/configuration_video_llava.py +0 -1
  1394. transformers/models/video_llava/image_processing_video_llava.py +35 -38
  1395. transformers/models/video_llava/modeling_video_llava.py +59 -57
  1396. transformers/models/video_llava/processing_video_llava.py +38 -78
  1397. transformers/models/video_llava/video_processing_video_llava.py +0 -1
  1398. transformers/models/videomae/configuration_videomae.py +0 -1
  1399. transformers/models/videomae/image_processing_videomae.py +31 -34
  1400. transformers/models/videomae/modeling_videomae.py +13 -15
  1401. transformers/models/videomae/video_processing_videomae.py +0 -1
  1402. transformers/models/vilt/configuration_vilt.py +2 -3
  1403. transformers/models/vilt/image_processing_vilt.py +29 -30
  1404. transformers/models/vilt/image_processing_vilt_fast.py +9 -10
  1405. transformers/models/vilt/modeling_vilt.py +83 -78
  1406. transformers/models/vilt/processing_vilt.py +2 -14
  1407. transformers/models/vipllava/configuration_vipllava.py +0 -1
  1408. transformers/models/vipllava/modeling_vipllava.py +45 -42
  1409. transformers/models/vipllava/modular_vipllava.py +30 -32
  1410. transformers/models/vision_encoder_decoder/configuration_vision_encoder_decoder.py +0 -1
  1411. transformers/models/vision_encoder_decoder/modeling_vision_encoder_decoder.py +18 -21
  1412. transformers/models/vision_text_dual_encoder/configuration_vision_text_dual_encoder.py +0 -1
  1413. transformers/models/vision_text_dual_encoder/modeling_vision_text_dual_encoder.py +18 -21
  1414. transformers/models/vision_text_dual_encoder/processing_vision_text_dual_encoder.py +2 -16
  1415. transformers/models/visual_bert/configuration_visual_bert.py +0 -1
  1416. transformers/models/visual_bert/modeling_visual_bert.py +92 -92
  1417. transformers/models/vit/configuration_vit.py +0 -1
  1418. transformers/models/vit/image_processing_vit.py +19 -22
  1419. transformers/models/vit/image_processing_vit_fast.py +0 -1
  1420. transformers/models/vit/modeling_vit.py +13 -15
  1421. transformers/models/vit_mae/configuration_vit_mae.py +0 -1
  1422. transformers/models/vit_mae/modeling_vit_mae.py +21 -23
  1423. transformers/models/vit_msn/configuration_vit_msn.py +0 -1
  1424. transformers/models/vit_msn/modeling_vit_msn.py +10 -12
  1425. transformers/models/vitdet/configuration_vitdet.py +0 -1
  1426. transformers/models/vitdet/modeling_vitdet.py +12 -14
  1427. transformers/models/vitmatte/configuration_vitmatte.py +2 -5
  1428. transformers/models/vitmatte/image_processing_vitmatte.py +15 -18
  1429. transformers/models/vitmatte/image_processing_vitmatte_fast.py +14 -16
  1430. transformers/models/vitmatte/modeling_vitmatte.py +13 -11
  1431. transformers/models/vitpose/configuration_vitpose.py +4 -7
  1432. transformers/models/vitpose/image_processing_vitpose.py +24 -25
  1433. transformers/models/vitpose/image_processing_vitpose_fast.py +9 -11
  1434. transformers/models/vitpose/modeling_vitpose.py +10 -12
  1435. transformers/models/vitpose_backbone/configuration_vitpose_backbone.py +0 -1
  1436. transformers/models/vitpose_backbone/modeling_vitpose_backbone.py +8 -10
  1437. transformers/models/vits/configuration_vits.py +0 -1
  1438. transformers/models/vits/modeling_vits.py +34 -35
  1439. transformers/models/vits/tokenization_vits.py +3 -4
  1440. transformers/models/vivit/configuration_vivit.py +0 -1
  1441. transformers/models/vivit/image_processing_vivit.py +36 -39
  1442. transformers/models/vivit/modeling_vivit.py +5 -7
  1443. transformers/models/vjepa2/__init__.py +0 -1
  1444. transformers/models/vjepa2/configuration_vjepa2.py +0 -1
  1445. transformers/models/vjepa2/modeling_vjepa2.py +30 -32
  1446. transformers/models/vjepa2/video_processing_vjepa2.py +0 -1
  1447. transformers/models/voxtral/__init__.py +0 -1
  1448. transformers/models/voxtral/configuration_voxtral.py +0 -1
  1449. transformers/models/voxtral/modeling_voxtral.py +19 -27
  1450. transformers/models/voxtral/modular_voxtral.py +12 -21
  1451. transformers/models/voxtral/processing_voxtral.py +25 -48
  1452. transformers/models/wav2vec2/configuration_wav2vec2.py +0 -1
  1453. transformers/models/wav2vec2/feature_extraction_wav2vec2.py +7 -10
  1454. transformers/models/wav2vec2/modeling_wav2vec2.py +67 -122
  1455. transformers/models/wav2vec2/processing_wav2vec2.py +6 -35
  1456. transformers/models/wav2vec2/tokenization_wav2vec2.py +20 -332
  1457. transformers/models/wav2vec2_bert/configuration_wav2vec2_bert.py +0 -1
  1458. transformers/models/wav2vec2_bert/modeling_wav2vec2_bert.py +65 -62
  1459. transformers/models/wav2vec2_bert/modular_wav2vec2_bert.py +52 -48
  1460. transformers/models/wav2vec2_bert/processing_wav2vec2_bert.py +6 -35
  1461. transformers/models/wav2vec2_conformer/configuration_wav2vec2_conformer.py +0 -1
  1462. transformers/models/wav2vec2_conformer/modeling_wav2vec2_conformer.py +84 -77
  1463. transformers/models/wav2vec2_conformer/modular_wav2vec2_conformer.py +37 -30
  1464. transformers/models/wav2vec2_phoneme/tokenization_wav2vec2_phoneme.py +16 -17
  1465. transformers/models/wav2vec2_with_lm/processing_wav2vec2_with_lm.py +36 -55
  1466. transformers/models/wavlm/configuration_wavlm.py +0 -1
  1467. transformers/models/wavlm/modeling_wavlm.py +45 -48
  1468. transformers/models/wavlm/modular_wavlm.py +4 -5
  1469. transformers/models/whisper/configuration_whisper.py +0 -1
  1470. transformers/models/whisper/english_normalizer.py +3 -4
  1471. transformers/models/whisper/feature_extraction_whisper.py +9 -24
  1472. transformers/models/whisper/generation_whisper.py +27 -48
  1473. transformers/models/whisper/modeling_whisper.py +73 -73
  1474. transformers/models/whisper/processing_whisper.py +3 -20
  1475. transformers/models/whisper/tokenization_whisper.py +9 -30
  1476. transformers/models/x_clip/configuration_x_clip.py +0 -1
  1477. transformers/models/x_clip/modeling_x_clip.py +70 -69
  1478. transformers/models/x_clip/processing_x_clip.py +2 -14
  1479. transformers/models/xcodec/configuration_xcodec.py +4 -6
  1480. transformers/models/xcodec/modeling_xcodec.py +20 -17
  1481. transformers/models/xglm/configuration_xglm.py +0 -1
  1482. transformers/models/xglm/modeling_xglm.py +59 -55
  1483. transformers/models/xglm/tokenization_xglm.py +1 -4
  1484. transformers/models/xlm/configuration_xlm.py +0 -1
  1485. transformers/models/xlm/modeling_xlm.py +139 -144
  1486. transformers/models/xlm/tokenization_xlm.py +3 -5
  1487. transformers/models/xlm_roberta/configuration_xlm_roberta.py +0 -1
  1488. transformers/models/xlm_roberta/modeling_xlm_roberta.py +195 -194
  1489. transformers/models/xlm_roberta/modular_xlm_roberta.py +50 -53
  1490. transformers/models/xlm_roberta/tokenization_xlm_roberta.py +1 -4
  1491. transformers/models/xlm_roberta_xl/configuration_xlm_roberta_xl.py +0 -1
  1492. transformers/models/xlm_roberta_xl/modeling_xlm_roberta_xl.py +94 -93
  1493. transformers/models/xlm_roberta_xl/modular_xlm_roberta_xl.py +67 -70
  1494. transformers/models/xlnet/configuration_xlnet.py +0 -11
  1495. transformers/models/xlnet/modeling_xlnet.py +152 -163
  1496. transformers/models/xlnet/tokenization_xlnet.py +1 -4
  1497. transformers/models/xlstm/configuration_xlstm.py +3 -5
  1498. transformers/models/xlstm/modeling_xlstm.py +62 -65
  1499. transformers/models/xmod/configuration_xmod.py +0 -1
  1500. transformers/models/xmod/modeling_xmod.py +101 -100
  1501. transformers/models/yolos/configuration_yolos.py +0 -1
  1502. transformers/models/yolos/image_processing_yolos.py +60 -62
  1503. transformers/models/yolos/image_processing_yolos_fast.py +18 -18
  1504. transformers/models/yolos/modeling_yolos.py +12 -14
  1505. transformers/models/yolos/modular_yolos.py +2 -4
  1506. transformers/models/yoso/configuration_yoso.py +0 -1
  1507. transformers/models/yoso/modeling_yoso.py +64 -63
  1508. transformers/models/zamba/configuration_zamba.py +0 -1
  1509. transformers/models/zamba/modeling_zamba.py +70 -70
  1510. transformers/models/zamba2/configuration_zamba2.py +36 -37
  1511. transformers/models/zamba2/modeling_zamba2.py +87 -89
  1512. transformers/models/zamba2/modular_zamba2.py +43 -45
  1513. transformers/models/zoedepth/configuration_zoedepth.py +1 -2
  1514. transformers/models/zoedepth/image_processing_zoedepth.py +28 -29
  1515. transformers/models/zoedepth/image_processing_zoedepth_fast.py +12 -15
  1516. transformers/models/zoedepth/modeling_zoedepth.py +21 -16
  1517. transformers/pipelines/__init__.py +59 -55
  1518. transformers/pipelines/any_to_any.py +14 -22
  1519. transformers/pipelines/audio_utils.py +1 -2
  1520. transformers/pipelines/automatic_speech_recognition.py +20 -12
  1521. transformers/pipelines/base.py +13 -17
  1522. transformers/pipelines/deprecated/__init__.py +0 -1
  1523. transformers/pipelines/document_question_answering.py +1 -1
  1524. transformers/pipelines/image_text_to_text.py +0 -1
  1525. transformers/pipelines/image_to_text.py +4 -44
  1526. transformers/pipelines/question_answering.py +5 -44
  1527. transformers/pipelines/text_classification.py +1 -14
  1528. transformers/pipelines/text_to_audio.py +2 -2
  1529. transformers/pipelines/token_classification.py +1 -22
  1530. transformers/pipelines/video_classification.py +1 -9
  1531. transformers/pipelines/zero_shot_audio_classification.py +0 -1
  1532. transformers/pipelines/zero_shot_classification.py +0 -6
  1533. transformers/pipelines/zero_shot_image_classification.py +0 -7
  1534. transformers/processing_utils.py +222 -151
  1535. transformers/quantizers/auto.py +2 -4
  1536. transformers/quantizers/base.py +19 -64
  1537. transformers/quantizers/quantizer_aqlm.py +1 -18
  1538. transformers/quantizers/quantizer_auto_round.py +1 -10
  1539. transformers/quantizers/quantizer_awq.py +3 -8
  1540. transformers/quantizers/quantizer_bitnet.py +1 -6
  1541. transformers/quantizers/quantizer_bnb_4bit.py +9 -49
  1542. transformers/quantizers/quantizer_bnb_8bit.py +9 -19
  1543. transformers/quantizers/quantizer_compressed_tensors.py +1 -4
  1544. transformers/quantizers/quantizer_eetq.py +2 -12
  1545. transformers/quantizers/quantizer_fbgemm_fp8.py +5 -14
  1546. transformers/quantizers/quantizer_finegrained_fp8.py +15 -10
  1547. transformers/quantizers/quantizer_fp_quant.py +4 -4
  1548. transformers/quantizers/quantizer_gptq.py +1 -4
  1549. transformers/quantizers/quantizer_higgs.py +2 -6
  1550. transformers/quantizers/quantizer_mxfp4.py +2 -28
  1551. transformers/quantizers/quantizer_quanto.py +14 -14
  1552. transformers/quantizers/quantizer_quark.py +0 -1
  1553. transformers/quantizers/quantizer_spqr.py +3 -8
  1554. transformers/quantizers/quantizer_torchao.py +31 -127
  1555. transformers/quantizers/quantizer_vptq.py +1 -10
  1556. transformers/testing_utils.py +31 -49
  1557. transformers/tokenization_mistral_common.py +554 -902
  1558. transformers/tokenization_utils_base.py +112 -124
  1559. transformers/tokenization_utils_sentencepiece.py +5 -6
  1560. transformers/tokenization_utils_tokenizers.py +30 -7
  1561. transformers/trainer.py +30 -11
  1562. transformers/trainer_callback.py +8 -0
  1563. transformers/trainer_jit_checkpoint.py +1 -2
  1564. transformers/trainer_seq2seq.py +4 -0
  1565. transformers/training_args.py +11 -13
  1566. transformers/utils/__init__.py +4 -0
  1567. transformers/utils/attention_visualizer.py +5 -5
  1568. transformers/utils/auto_docstring.py +598 -37
  1569. transformers/utils/doc.py +1 -1
  1570. transformers/utils/dummy_pt_objects.py +0 -42
  1571. transformers/utils/generic.py +21 -1
  1572. transformers/utils/import_utils.py +51 -9
  1573. transformers/utils/kernel_config.py +71 -18
  1574. transformers/utils/loading_report.py +3 -3
  1575. transformers/utils/quantization_config.py +16 -18
  1576. transformers/video_processing_utils.py +35 -32
  1577. transformers/video_utils.py +18 -22
  1578. {transformers-5.0.0rc1.dist-info → transformers-5.0.0rc3.dist-info}/METADATA +23 -24
  1579. transformers-5.0.0rc3.dist-info/RECORD +2067 -0
  1580. transformers-5.0.0rc1.dist-info/RECORD +0 -2003
  1581. {transformers-5.0.0rc1.dist-info → transformers-5.0.0rc3.dist-info}/WHEEL +0 -0
  1582. {transformers-5.0.0rc1.dist-info → transformers-5.0.0rc3.dist-info}/entry_points.txt +0 -0
  1583. {transformers-5.0.0rc1.dist-info → transformers-5.0.0rc3.dist-info}/licenses/LICENSE +0 -0
  1584. {transformers-5.0.0rc1.dist-info → transformers-5.0.0rc3.dist-info}/top_level.txt +0 -0
transformers/modeling_outputs.py CHANGED
@@ -12,9 +12,7 @@
12
12
  # See the License for the specific language governing permissions and
13
13
  # limitations under the License.
14
14
 
15
- import warnings
16
15
  from dataclasses import dataclass
17
- from typing import Optional
18
16
 
19
17
  import torch
20
18
 
@@ -43,9 +41,9 @@ class BaseModelOutput(ModelOutput):
43
41
  heads.
44
42
  """
45
43
 
46
- last_hidden_state: Optional[torch.FloatTensor] = None
47
- hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
48
- attentions: Optional[tuple[torch.FloatTensor, ...]] = None
44
+ last_hidden_state: torch.FloatTensor | None = None
45
+ hidden_states: tuple[torch.FloatTensor, ...] | None = None
46
+ attentions: tuple[torch.FloatTensor, ...] | None = None
49
47
 
50
48
 
51
49
  @dataclass
@@ -63,8 +61,8 @@ class BaseModelOutputWithNoAttention(ModelOutput):
63
61
  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
64
62
  """
65
63
 
66
- last_hidden_state: Optional[torch.FloatTensor] = None
67
- hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
64
+ last_hidden_state: torch.FloatTensor | None = None
65
+ hidden_states: tuple[torch.FloatTensor, ...] | None = None
68
66
 
69
67
 
70
68
  @dataclass
@@ -93,10 +91,10 @@ class BaseModelOutputWithPooling(ModelOutput):
93
91
  heads.
94
92
  """
95
93
 
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
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
100
98
 
101
99
 
102
100
  @dataclass
@@ -116,9 +114,9 @@ class BaseModelOutputWithPoolingAndNoAttention(ModelOutput):
116
114
  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
117
115
  """
118
116
 
119
- last_hidden_state: Optional[torch.FloatTensor] = None
120
- pooler_output: Optional[torch.FloatTensor] = None
121
- hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
117
+ last_hidden_state: torch.FloatTensor | None = None
118
+ pooler_output: torch.FloatTensor | None = None
119
+ hidden_states: tuple[torch.FloatTensor, ...] | None = None
122
120
 
123
121
 
124
122
  @dataclass
@@ -151,10 +149,10 @@ class BaseModelOutputWithPast(ModelOutput):
151
149
  heads.
152
150
  """
153
151
 
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
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
158
156
 
159
157
 
160
158
  @dataclass
@@ -184,10 +182,10 @@ class BaseModelOutputWithCrossAttentions(ModelOutput):
184
182
  weighted average in the cross-attention heads.
185
183
  """
186
184
 
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
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
191
189
 
192
190
 
193
191
  @dataclass
@@ -228,12 +226,12 @@ class BaseModelOutputWithPoolingAndCrossAttentions(ModelOutput):
228
226
  input) to speed up sequential decoding.
229
227
  """
230
228
 
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
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
237
235
 
238
236
 
239
237
  @dataclass
@@ -272,11 +270,11 @@ class BaseModelOutputWithPastAndCrossAttentions(ModelOutput):
272
270
  weighted average in the cross-attention heads.
273
271
  """
274
272
 
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
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
280
278
 
281
279
 
282
280
  @dataclass
@@ -317,14 +315,14 @@ class MoECausalLMOutputWithPast(ModelOutput):
317
315
  modules.
318
316
  """
319
317
 
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
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
328
326
 
329
327
 
330
328
  @dataclass
@@ -353,11 +351,11 @@ class MoEModelOutput(ModelOutput):
353
351
  loss and the z_loss for Mixture of Experts models.
354
352
  """
355
353
 
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
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
361
359
 
362
360
 
363
361
  @dataclass
@@ -392,11 +390,11 @@ class MoeModelOutputWithPast(ModelOutput):
392
390
  loss for Mixture of Experts models.
393
391
  """
394
392
 
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
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
400
398
 
401
399
 
402
400
  @dataclass
@@ -438,13 +436,13 @@ class MoeCausalLMOutputWithPast(ModelOutput):
438
436
  heads.
439
437
  """
440
438
 
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
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
448
446
 
449
447
 
450
448
  @dataclass
@@ -489,13 +487,13 @@ class MoEModelOutputWithPastAndCrossAttentions(ModelOutput):
489
487
  loss and the z_loss for Mixture of Experts models.
490
488
  """
491
489
 
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
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
499
497
 
500
498
 
501
499
  @dataclass
@@ -547,14 +545,14 @@ class Seq2SeqModelOutput(ModelOutput):
547
545
  self-attention heads.
548
546
  """
549
547
 
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
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
558
556
 
559
557
 
560
558
  @dataclass
@@ -615,16 +613,16 @@ class Seq2SeqMoEModelOutput(ModelOutput):
615
613
  modules.
616
614
  """
617
615
 
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
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
628
626
 
629
627
 
630
628
  @dataclass
@@ -650,10 +648,10 @@ class CausalLMOutput(ModelOutput):
650
648
  heads.
651
649
  """
652
650
 
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
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
657
655
 
658
656
 
659
657
  @dataclass
@@ -684,11 +682,11 @@ class CausalLMOutputWithPast(ModelOutput):
684
682
  heads.
685
683
  """
686
684
 
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
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
692
690
 
693
691
 
694
692
  @dataclass
@@ -725,12 +723,12 @@ class CausalLMOutputWithCrossAttentions(ModelOutput):
725
723
  `past_key_values` input) to speed up sequential decoding.
726
724
  """
727
725
 
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
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
734
732
 
735
733
 
736
734
  @dataclass
@@ -761,11 +759,11 @@ class SequenceClassifierOutputWithPast(ModelOutput):
761
759
  heads.
762
760
  """
763
761
 
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
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
769
767
 
770
768
 
771
769
  @dataclass
@@ -791,10 +789,10 @@ class MaskedLMOutput(ModelOutput):
791
789
  heads.
792
790
  """
793
791
 
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
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
798
796
 
799
797
 
800
798
  @dataclass
@@ -844,15 +842,15 @@ class Seq2SeqLMOutput(ModelOutput):
844
842
  self-attention heads.
845
843
  """
846
844
 
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
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
856
854
 
857
855
 
858
856
  @dataclass
@@ -911,21 +909,21 @@ class Seq2SeqMoEOutput(ModelOutput):
911
909
  models.
912
910
  """
913
911
 
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
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
929
927
 
930
928
 
931
929
  @dataclass
@@ -952,10 +950,10 @@ class NextSentencePredictorOutput(ModelOutput):
952
950
  heads.
953
951
  """
954
952
 
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
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
959
957
 
960
958
 
961
959
  @dataclass
@@ -981,10 +979,10 @@ class SequenceClassifierOutput(ModelOutput):
981
979
  heads.
982
980
  """
983
981
 
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
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
988
986
 
989
987
 
990
988
  @dataclass
@@ -1034,15 +1032,15 @@ class Seq2SeqSequenceClassifierOutput(ModelOutput):
1034
1032
  self-attention heads.
1035
1033
  """
1036
1034
 
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
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
1046
1044
 
1047
1045
 
1048
1046
  @dataclass
@@ -1070,10 +1068,10 @@ class MultipleChoiceModelOutput(ModelOutput):
1070
1068
  heads.
1071
1069
  """
1072
1070
 
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
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
1077
1075
 
1078
1076
 
1079
1077
  @dataclass
@@ -1099,10 +1097,10 @@ class TokenClassifierOutput(ModelOutput):
1099
1097
  heads.
1100
1098
  """
1101
1099
 
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
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
1106
1104
 
1107
1105
 
1108
1106
  @dataclass
@@ -1130,11 +1128,11 @@ class QuestionAnsweringModelOutput(ModelOutput):
1130
1128
  heads.
1131
1129
  """
1132
1130
 
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
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
1138
1136
 
1139
1137
 
1140
1138
  @dataclass
@@ -1186,16 +1184,16 @@ class Seq2SeqQuestionAnsweringModelOutput(ModelOutput):
1186
1184
  self-attention heads.
1187
1185
  """
1188
1186
 
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
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
1199
1197
 
1200
1198
 
1201
1199
  @dataclass
@@ -1230,10 +1228,10 @@ class SemanticSegmenterOutput(ModelOutput):
1230
1228
  heads.
1231
1229
  """
1232
1230
 
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
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
1237
1235
 
1238
1236
 
1239
1237
  @dataclass
@@ -1258,10 +1256,10 @@ class ImageClassifierOutput(ModelOutput):
1258
1256
  heads.
1259
1257
  """
1260
1258
 
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
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
1265
1263
 
1266
1264
 
1267
1265
  @dataclass
@@ -1280,9 +1278,9 @@ class ImageClassifierOutputWithNoAttention(ModelOutput):
1280
1278
  called feature maps) of the model at the output of each stage.
1281
1279
  """
1282
1280
 
1283
- loss: Optional[torch.FloatTensor] = None
1284
- logits: Optional[torch.FloatTensor] = None
1285
- hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1281
+ loss: torch.FloatTensor | None = None
1282
+ logits: torch.FloatTensor | None = None
1283
+ hidden_states: tuple[torch.FloatTensor, ...] | None = None
1286
1284
 
1287
1285
 
1288
1286
  @dataclass
@@ -1309,10 +1307,10 @@ class DepthEstimatorOutput(ModelOutput):
1309
1307
  heads.
1310
1308
  """
1311
1309
 
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
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
1316
1314
 
1317
1315
 
1318
1316
  @dataclass
@@ -1337,10 +1335,10 @@ class ImageSuperResolutionOutput(ModelOutput):
1337
1335
  heads.
1338
1336
  """
1339
1337
 
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
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
1344
1342
 
1345
1343
 
1346
1344
  @dataclass
@@ -1366,10 +1364,10 @@ class Wav2Vec2BaseModelOutput(ModelOutput):
1366
1364
  heads.
1367
1365
  """
1368
1366
 
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
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
1373
1371
 
1374
1372
 
1375
1373
  @dataclass
@@ -1397,11 +1395,11 @@ class XVectorOutput(ModelOutput):
1397
1395
  heads.
1398
1396
  """
1399
1397
 
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
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
1405
1403
 
1406
1404
 
1407
1405
  @dataclass
@@ -1426,9 +1424,9 @@ class BackboneOutput(ModelOutput):
1426
1424
  heads.
1427
1425
  """
1428
1426
 
1429
- feature_maps: Optional[tuple[torch.FloatTensor]] = None
1430
- hidden_states: Optional[tuple[torch.FloatTensor, ...]] = None
1431
- attentions: Optional[tuple[torch.FloatTensor, ...]] = None
1427
+ feature_maps: tuple[torch.FloatTensor] | None = None
1428
+ hidden_states: tuple[torch.FloatTensor, ...] | None = None
1429
+ attentions: tuple[torch.FloatTensor, ...] | None = None
1432
1430
 
1433
1431
 
1434
1432
  @dataclass
@@ -1461,11 +1459,11 @@ class BaseModelOutputWithPoolingAndProjection(ModelOutput):
1461
1459
  Text embeddings before the projection layer, used to mimic the last hidden state of the teacher encoder.
1462
1460
  """
1463
1461
 
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
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
1469
1467
 
1470
1468
 
1471
1469
  @dataclass
@@ -1515,15 +1513,15 @@ class Seq2SeqSpectrogramOutput(ModelOutput):
1515
1513
  self-attention heads.
1516
1514
  """
1517
1515
 
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
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
1527
1525
 
1528
1526
 
1529
1527
  @dataclass
@@ -1583,17 +1581,17 @@ class Seq2SeqTSModelOutput(ModelOutput):
1583
1581
  Static features of each time series' in a batch which are copied to the covariates at inference time.
1584
1582
  """
1585
1583
 
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
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
1597
1595
 
1598
1596
 
1599
1597
  @dataclass
@@ -1652,18 +1650,18 @@ class Seq2SeqTSPredictionOutput(ModelOutput):
1652
1650
  Static features of each time series' in a batch which are copied to the covariates at inference time.
1653
1651
  """
1654
1652
 
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
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
1667
1665
 
1668
1666
 
1669
1667
  @dataclass
@@ -1677,7 +1675,7 @@ class SampleTSPredictionOutput(ModelOutput):
1677
1675
  Sampled values from the chosen distribution.
1678
1676
  """
1679
1677
 
1680
- sequences: Optional[torch.FloatTensor] = None
1678
+ sequences: torch.FloatTensor | None = None
1681
1679
 
1682
1680
 
1683
1681
  @dataclass
@@ -1702,16 +1700,7 @@ class MaskedImageModelingOutput(ModelOutput):
1702
1700
  the self-attention heads.
1703
1701
  """
1704
1702
 
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
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