liger-kernel-nightly 0.6.1.dev20250812205818__py3-none-any.whl → 0.6.1.dev20250819145841__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 (7) hide show
  1. liger_kernel/transformers/model/mllama.py +4 -2
  2. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/METADATA +1 -1
  3. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/RECORD +7 -7
  4. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/LICENSE +0 -0
  5. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/NOTICE +0 -0
  6. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/WHEEL +0 -0
  7. {liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/top_level.txt +0 -0
liger_kernel/transformers/model/mllama.py CHANGED
@@ -190,7 +190,9 @@ def lce_forward(
190
190
  output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
191
191
  )
192
192
  return_dict = return_dict if return_dict is not None else self.config.use_return_dict
193
-
193
+ # Filter out accum_dtype from kwargs for model call as MllamaTextModel doesn't accept it in transformers 4.49.0
194
+ # but preserve it for loss function calls
195
+ model_kwargs = {k: v for k, v in kwargs.items() if k != "accum_dtype"}
194
196
  # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
195
197
  outputs = self.model(
196
198
  input_ids=input_ids,
@@ -206,7 +208,7 @@ def lce_forward(
206
208
  output_hidden_states=output_hidden_states,
207
209
  return_dict=return_dict,
208
210
  cache_position=cache_position,
209
- **kwargs,
211
+ **model_kwargs,
210
212
  )
211
213
 
212
214
  hidden_states = outputs[0]
{liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: liger_kernel_nightly
3
- Version: 0.6.1.dev20250812205818
3
+ Version: 0.6.1.dev20250819145841
4
4
  Summary: Efficient Triton kernels for LLM Training
5
5
  License: BSD 2-CLAUSE LICENSE
6
6
  Copyright 2024 LinkedIn Corporation
{liger_kernel_nightly-0.6.1.dev20250812205818.dist-info → liger_kernel_nightly-0.6.1.dev20250819145841.dist-info}/RECORD RENAMED
@@ -80,7 +80,7 @@ liger_kernel/transformers/model/llava.py,sha256=bLCioday_SOm69ogMDBhy_4UsVkH2-BS
80
80
  liger_kernel/transformers/model/loss_utils.py,sha256=YiYsmRHIuoRnFjGpwyIM18DCsrPPmO32YWMWqkEm1UQ,1867
81
81
  liger_kernel/transformers/model/mistral.py,sha256=syYNL8dLThX2-4uC13Lu0krEZ5zw3InviDUR3AJmc-I,5500
82
82
  liger_kernel/transformers/model/mixtral.py,sha256=VY-y73IyjcCyWyI7ahxXLw0fJrhgjYfr1xwRYtsHX0o,11396
83
- liger_kernel/transformers/model/mllama.py,sha256=my29NXk-p6ckQaP8qDIN8e318yI_9mQZHt38MV3SqLY,11280
83
+ liger_kernel/transformers/model/mllama.py,sha256=NhJtlXiuszJHo5YSJOvSGYH47ly7Hse8r-5BKznBg9s,11522
84
84
  liger_kernel/transformers/model/olmo2.py,sha256=6L_bo-ZUgO1lYppdJneOtYxNIylQKS6BiGp13g7Uq9E,5259
85
85
  liger_kernel/transformers/model/paligemma.py,sha256=xuIx3oOwTgftU3jqLfWOxUxgCLBNJh0yNC21an9qDjo,18773
86
86
  liger_kernel/transformers/model/phi3.py,sha256=AwScxUe3LjmHHyQg4gW9bMoUI7uA6fUEMXJ3YhBiHtQ,4046
@@ -94,9 +94,9 @@ liger_kernel/transformers/trainer/__init__.py,sha256=p7yQfklV8-467qSz_ZMimkbDF7H
94
94
  liger_kernel/transformers/trainer/orpo_trainer.py,sha256=tX0h63aOFe3rNqTmk6JpMf75UPo981yzEa6TghnjS0Q,5370
95
95
  liger_kernel/triton/__init__.py,sha256=qCiCamzCRv6lpV8IqpAc9YMdNKC7GKurClWceQPnlis,92
96
96
  liger_kernel/triton/monkey_patch.py,sha256=Rd0hUHAzDkFfHvnX7-PBaNK5EKnZhtfM_h-fgQH9HPY,1568
97
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/LICENSE,sha256=OhzLDHJ0to4a8sodVLELZiCFylZ1NAAYLs-HrjPy0ag,1312
98
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/METADATA,sha256=1s-igyDBWH7I09Q1f-7-h5BtStzEN30M_ffGQk5ZE4M,24504
99
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/NOTICE,sha256=njwnoPZLh9AN8SJQzxvCGLHi-8X__AvWRze6joNXIY8,2066
100
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/WHEEL,sha256=iAkIy5fosb7FzIOwONchHf19Qu7_1wCWyFNR5gu9nU0,91
101
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/top_level.txt,sha256=2eghu4hA3LnkM7ElW92tQ8zegWKgSbeo-k-aGe1YnvY,13
102
- liger_kernel_nightly-0.6.1.dev20250812205818.dist-info/RECORD,,
97
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/LICENSE,sha256=OhzLDHJ0to4a8sodVLELZiCFylZ1NAAYLs-HrjPy0ag,1312
98
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/METADATA,sha256=h8wVtgQPQH2HPJAlAIPPzrkbR0YlZqpkKBLylhJXQdo,24504
99
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/NOTICE,sha256=njwnoPZLh9AN8SJQzxvCGLHi-8X__AvWRze6joNXIY8,2066
100
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/WHEEL,sha256=iAkIy5fosb7FzIOwONchHf19Qu7_1wCWyFNR5gu9nU0,91
101
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/top_level.txt,sha256=2eghu4hA3LnkM7ElW92tQ8zegWKgSbeo-k-aGe1YnvY,13
102
+ liger_kernel_nightly-0.6.1.dev20250819145841.dist-info/RECORD,,