project-llm-trainer 0.4.3__py3-none-any.whl → 0.4.4__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.

Potentially problematic release.

This version of project-llm-trainer might be problematic. Click here for more details.

Files changed (12) hide show
  1. llm_trainer/trainer.py +4 -4
  2. {project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/METADATA +1 -1
  3. {project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/RECORD +12 -12
  4. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/calc_intermediate_size +0 -0
  5. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/ddp_train +0 -0
  6. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/ds_train +0 -0
  7. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/plot_loss +0 -0
  8. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/plot_lr +0 -0
  9. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/py_train +0 -0
  10. {project_llm_trainer-0.4.3.data → project_llm_trainer-0.4.4.data}/scripts/smart_train +0 -0
  11. {project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/WHEEL +0 -0
  12. {project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/top_level.txt +0 -0
llm_trainer/trainer.py CHANGED
@@ -139,10 +139,10 @@ class Trainer:
139
139
  if not any(sub_module in name for sub_module in ['multi_modal_projector']):
140
140
  param.requires_grad = False
141
141
 
142
- model.embed_tokens.eval()
143
- model.layers.eval()
144
- model.head_norm.eval()
145
- model.lm_head.eval()
142
+ # model.embed_tokens.eval()
143
+ # model.layers.eval()
144
+ # model.head_norm.eval()
145
+ # model.lm_head.eval()
146
146
 
147
147
  if TrainerTools().parallel.is_main_process:
148
148
  total_params = sum(p.numel() for p in model.parameters())
{project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: project_llm_trainer
3
- Version: 0.4.3
3
+ Version: 0.4.4
4
4
  Summary: LLM and VLM trainer
5
5
  Author: qibin
6
6
  Author-email: qibin0506@gmail.com
{project_llm_trainer-0.4.3.dist-info → project_llm_trainer-0.4.4.dist-info}/RECORD RENAMED
@@ -20,16 +20,16 @@ llm_trainer/sft_trainer.py,sha256=gxQA7T1o1QGUsHp2CX1Qb_fO5LppBJuNbc0H4ixCYUA,17
20
20
  llm_trainer/tokenizer.py,sha256=A7TYYUbtPf75kjCvWP7yBui4xZBObMk2aPem62YpwpY,6776
21
21
  llm_trainer/tools.py,sha256=O45-20wRmh-nyTfU-U-XtjbKAoe7boEIsUvWT_NaKx4,3041
22
22
  llm_trainer/train_configs.py,sha256=arnet3tIzgVnwshod08F1jE7r4I7e-SIgMy55IagPnE,15971
23
- llm_trainer/trainer.py,sha256=aoZYL5U4Z5axXBMM_DHgzIzJ89YbU9xUQ56jppcT65c,25339
23
+ llm_trainer/trainer.py,sha256=Aynp-lT1-lHHJ2S7FboHb57M4bPe51lLSgM6SOICQOo,25347
24
24
  llm_trainer/utils.py,sha256=-ivhMF0d999va13S1wt2uBvtVw8Nvr3uBzhaUFKL04Q,6826
25
- project_llm_trainer-0.4.3.data/scripts/calc_intermediate_size,sha256=AggpgNHokJiJMbEtVdOnolqr_4bH3i1UYuZNEAzC2Gc,460
26
- project_llm_trainer-0.4.3.data/scripts/ddp_train,sha256=x81AasaN2-9TwARFFF1l7iV1LmfMQ0bLw0i_CGbOwSw,299
27
- project_llm_trainer-0.4.3.data/scripts/ds_train,sha256=qL3qc3TcedBCw98UZUjW07ONcErRawLE1HymW2AmscA,265
28
- project_llm_trainer-0.4.3.data/scripts/plot_loss,sha256=MzFcdJESlVr1srj4Td6-AxPGUKkfB_QEcJwm0Bd-5fU,910
29
- project_llm_trainer-0.4.3.data/scripts/plot_lr,sha256=w_7XR_x3KYYyboeOVAeu_I4fveLFI-C0wBmRrNlmWUI,894
30
- project_llm_trainer-0.4.3.data/scripts/py_train,sha256=tOp9TquORQeU8XN5H7OVIk5O0Ypwi34p_GENxTwgwdk,265
31
- project_llm_trainer-0.4.3.data/scripts/smart_train,sha256=Pmt4Q0to4Hoz82iB9uFPZuz7uahNUbfE7FR1940EBy8,716
32
- project_llm_trainer-0.4.3.dist-info/METADATA,sha256=kmmc6L6SE9iBvNHutWpeb0TocGX5vixhvHHLS4ltqec,195
33
- project_llm_trainer-0.4.3.dist-info/WHEEL,sha256=Nw36Djuh_5VDukK0H78QzOX-_FQEo6V37m3nkm96gtU,91
34
- project_llm_trainer-0.4.3.dist-info/top_level.txt,sha256=LtRFg28i0QIG7iBCD2t095oSco99LCtkijibS9cMGik,12
35
- project_llm_trainer-0.4.3.dist-info/RECORD,,
25
+ project_llm_trainer-0.4.4.data/scripts/calc_intermediate_size,sha256=AggpgNHokJiJMbEtVdOnolqr_4bH3i1UYuZNEAzC2Gc,460
26
+ project_llm_trainer-0.4.4.data/scripts/ddp_train,sha256=x81AasaN2-9TwARFFF1l7iV1LmfMQ0bLw0i_CGbOwSw,299
27
+ project_llm_trainer-0.4.4.data/scripts/ds_train,sha256=qL3qc3TcedBCw98UZUjW07ONcErRawLE1HymW2AmscA,265
28
+ project_llm_trainer-0.4.4.data/scripts/plot_loss,sha256=MzFcdJESlVr1srj4Td6-AxPGUKkfB_QEcJwm0Bd-5fU,910
29
+ project_llm_trainer-0.4.4.data/scripts/plot_lr,sha256=w_7XR_x3KYYyboeOVAeu_I4fveLFI-C0wBmRrNlmWUI,894
30
+ project_llm_trainer-0.4.4.data/scripts/py_train,sha256=tOp9TquORQeU8XN5H7OVIk5O0Ypwi34p_GENxTwgwdk,265
31
+ project_llm_trainer-0.4.4.data/scripts/smart_train,sha256=Pmt4Q0to4Hoz82iB9uFPZuz7uahNUbfE7FR1940EBy8,716
32
+ project_llm_trainer-0.4.4.dist-info/METADATA,sha256=7cpmCfNcYyufr2GgoiK3UoelmgWQZh5BOi6rCXrxxkA,195
33
+ project_llm_trainer-0.4.4.dist-info/WHEEL,sha256=Nw36Djuh_5VDukK0H78QzOX-_FQEo6V37m3nkm96gtU,91
34
+ project_llm_trainer-0.4.4.dist-info/top_level.txt,sha256=LtRFg28i0QIG7iBCD2t095oSco99LCtkijibS9cMGik,12
35
+ project_llm_trainer-0.4.4.dist-info/RECORD,,