Daniel Han
74382dea47
Update rms_layernorm.py
2024-11-16 12:18:47 -08:00
Edd
a146521a50
fix/sfttrainer-compatibility ( #1293 )
...
* Refactor trainer.py to import SFTConfig directly and update UnslothTrainingArguments class inheritance
* Update trainer.py
* Update trainer.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-14 17:07:29 -08:00
Daniel Han
5eb971fa4e
Fix #853
2024-11-14 01:26:13 -08:00
Daniel Han
3b11ae7dd2
Update llama.py
2024-11-14 01:11:16 -08:00
Daniel Han
e4d1754c0b
Merge branch 'main' into nightly
2024-11-13 19:07:33 -08:00
Daniel Han
f26d4e739e
Update _utils.py
2024-11-13 19:07:26 -08:00
Daniel Han
0de5457252
Bug fixes ( #1288 )
...
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
* Update _utils.py
* fix/transformers-unpack (#1180 )
* Fix DPO, ORPO (#1177 )
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Add warning for missing Unpack and KwargsForCausalLM in older Transformers versions
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Update cross_entropy_loss.py
* Update _utils.py
* Update _utils.py
* donot upcast lm_head and embeddings to float32 (#1186 )
* Cleanup upcast logs (#1188 )
* Fix/phi-longrope (#1193 )
* Enhance rotary embedding handling in LlamaAttention and LongRopeRotaryEmbedding
* Typo
* Improve rotary embedding handling in LlamaAttention to prevent errors with short KV cache
* Update llama.py
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Update transformers
* Unk token issues
* Update _utils.py
* Fix pad token
* Update llama.py
* Typo
* ignored labels
* Revert "ignored labels"
This reverts commit 9d07be077b#1219 )
* Update save.py
Check whether path is in /tmp dir for Kaggle environment
* Update save.py
Move temporary_location to /tmp in Kaggle
* Enhance Kaggle environment support in save and tokenizer utilities
---------
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
* Bug fixes
* Update pyproject.toml
* Update _utils.py
* Update __init__.py
* Update __init__.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Tied weights
* Revert "Tied weights"
This reverts commit 8090b7c01a#1254 )
* Fix: cast logits to float32 in cross_entropy_forward to prevent errors
* Update cross_entropy_loss.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Throw error when inferencing longer than max_popsition_embeddings (#1236 )
* Throw error when inferencing longer than max_popsition_embeddings without rope scaling
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* CLI now handles user input strings for dtype correctly (#1235 )
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
* Update flex_attention.py
* Update _utils.py
* Update _utils.py
* Update flex_attention.py
* Update flex_attention.py
* Update loader.py
* Update loader.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update _utils.py
* Update cross_entropy_loss.py
* Update _utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* triton_cast
* Update utils.py
* Qwen 2.5 Coder
* Fix/export mistral (#1281 )
* Enhance install_python_non_blocking to handle protobuf installation and process management
* Revert "Enhance install_python_non_blocking to handle protobuf installation and process management"
This reverts commit f09974b151#1266
* Revert "Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION to 'python' to address issue #1266 "
This reverts commit 9fc130785d#1266
* Update __init__.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* DOC Update - Update README.md with os.environ in example (#1269 )
* Update README.md with os.environ in example
Added OS Environ in example to avoid device conflicts , for a user at least in jupyter notebook this allows to select GPU in a multi GPU setup.
As currently the unsloth init checks all GPU's and takes the first in the order which can be a issue when some GPU's are in use and the list still shows them. So to manually avoid this, this os config is required.
Small change but a bit time saver for those who straight away copies the tutorials
* Update README.md
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* fix/get_chat_template (#1246 )
* Refactor `get_chat_template` to now support system message instead. It supposed to fix ollama tokenizer chattemplate to
* Remove type hinting
* Update chat_templates.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* fix/sft-trainer (#1276 )
* Add patch for SFTTrainer to maintain backward compatibility with TRL changes
* Update trainer.py
* Update trainer.py
* Refactor trainer patch to maintain backward compatibility with TRL changes
* Update trainer.py
* Refactor trainer.py to exclude non-convertible trainers from backward compatibility patch
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Update __init__.py
* Update trainer.py
* Update trainer.py
* Update trainer.py
* Update tokenizer_utils.py
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
Co-authored-by: Edd <68678137+Erland366@users.noreply.github.com>
Co-authored-by: Datta Nimmaturi <datta.nimmaturi@nutanix.com>
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
Co-authored-by: Edwin Fennell <edwinfennell1@gmail.com>
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
Co-authored-by: Uday Girish Maradana <einsteingirish@gmail.com>
2024-11-13 19:05:40 -08:00
Daniel Han
11b350f7c1
Update tokenizer_utils.py
2024-11-13 19:05:15 -08:00
Daniel Han
597169c14b
Update trainer.py
2024-11-13 18:53:40 -08:00
Daniel Han
035bccee75
Update trainer.py
2024-11-13 18:48:59 -08:00
Daniel Han
a31027c9e4
Update trainer.py
2024-11-13 18:44:54 -08:00
Daniel Han
84d6d36cde
Update __init__.py
2024-11-13 17:38:26 -08:00
Edd
dc0232c883
fix/sft-trainer ( #1276 )
...
* Add patch for SFTTrainer to maintain backward compatibility with TRL changes
* Update trainer.py
* Update trainer.py
* Refactor trainer patch to maintain backward compatibility with TRL changes
* Update trainer.py
* Refactor trainer.py to exclude non-convertible trainers from backward compatibility patch
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-13 17:33:30 -08:00
Edd
10565efe27
fix/get_chat_template ( #1246 )
...
* Refactor `get_chat_template` to now support system message instead. It supposed to fix ollama tokenizer chattemplate to
* Remove type hinting
* Update chat_templates.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-13 00:06:48 -08:00
Uday Girish Maradana
03c624375f
DOC Update - Update README.md with os.environ in example ( #1269 )
...
* Update README.md with os.environ in example
Added OS Environ in example to avoid device conflicts , for a user at least in jupyter notebook this allows to select GPU in a multi GPU setup.
As currently the unsloth init checks all GPU's and takes the first in the order which can be a issue when some GPU's are in use and the list still shows them. So to manually avoid this, this os config is required.
Small change but a bit time saver for those who straight away copies the tutorials
* Update README.md
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-12 23:55:28 -08:00
Edd
3fec577bb2
Fix/export mistral ( #1281 )
...
* Enhance install_python_non_blocking to handle protobuf installation and process management
* Revert "Enhance install_python_non_blocking to handle protobuf installation and process management"
This reverts commit f09974b151#1266
* Revert "Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION to 'python' to address issue #1266 "
This reverts commit 9fc130785d#1266
* Update __init__.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-12 23:53:50 -08:00
Daniel Han
e99acdd854
Merge branch 'main' into nightly
2024-11-12 23:51:46 -08:00
Daniel Han
d8ff860c84
Update _utils.py
2024-11-12 10:54:58 -08:00
Daniel Han
899caf0bb5
Qwen 2.5 ( #1280 )
...
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
* Update _utils.py
* fix/transformers-unpack (#1180 )
* Fix DPO, ORPO (#1177 )
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Add warning for missing Unpack and KwargsForCausalLM in older Transformers versions
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Update cross_entropy_loss.py
* Update _utils.py
* Update _utils.py
* donot upcast lm_head and embeddings to float32 (#1186 )
* Cleanup upcast logs (#1188 )
* Fix/phi-longrope (#1193 )
* Enhance rotary embedding handling in LlamaAttention and LongRopeRotaryEmbedding
* Typo
* Improve rotary embedding handling in LlamaAttention to prevent errors with short KV cache
* Update llama.py
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Update transformers
* Unk token issues
* Update _utils.py
* Fix pad token
* Update llama.py
* Typo
* ignored labels
* Revert "ignored labels"
This reverts commit 9d07be077b#1219 )
* Update save.py
Check whether path is in /tmp dir for Kaggle environment
* Update save.py
Move temporary_location to /tmp in Kaggle
* Enhance Kaggle environment support in save and tokenizer utilities
---------
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
* Bug fixes
* Update pyproject.toml
* Update _utils.py
* Update __init__.py
* Update __init__.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Tied weights
* Revert "Tied weights"
This reverts commit 8090b7c01a#1254 )
* Fix: cast logits to float32 in cross_entropy_forward to prevent errors
* Update cross_entropy_loss.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Throw error when inferencing longer than max_popsition_embeddings (#1236 )
* Throw error when inferencing longer than max_popsition_embeddings without rope scaling
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* CLI now handles user input strings for dtype correctly (#1235 )
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
* Update flex_attention.py
* Update _utils.py
* Update _utils.py
* Update flex_attention.py
* Update flex_attention.py
* Update loader.py
* Update loader.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update _utils.py
* Update cross_entropy_loss.py
* Update _utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* triton_cast
* Update utils.py
* Qwen 2.5 Coder
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
Co-authored-by: Edd <68678137+Erland366@users.noreply.github.com>
Co-authored-by: Datta Nimmaturi <datta.nimmaturi@nutanix.com>
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
Co-authored-by: Edwin Fennell <edwinfennell1@gmail.com>
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
2024-11-12 03:22:41 -08:00
Daniel Han
af9ba073da
Qwen 2.5 Coder
2024-11-11 18:46:05 -08:00
Daniel Han
fcf200997a
Update utils.py
2024-11-11 00:37:37 -08:00
Daniel Han
ad41479c54
triton_cast
2024-11-11 00:17:22 -08:00
Daniel Han
8fe9109431
Update tokenizer_utils.py
2024-11-11 00:04:02 -08:00
Daniel Han
673f541788
Update tokenizer_utils.py
2024-11-09 17:40:32 -08:00
Daniel Han
92c6a27840
Update tokenizer_utils.py
2024-11-09 17:37:11 -08:00
Daniel Han
a4e9705531
Update tokenizer_utils.py
2024-11-09 17:34:47 -08:00
Daniel Han
f1ec165096
Update tokenizer_utils.py
2024-11-09 17:01:33 -08:00
Daniel Han
3a1e7ef829
Update _utils.py
2024-11-07 01:11:45 -08:00
Daniel Han
8b3e9c2ff0
Update cross_entropy_loss.py
2024-11-06 21:07:50 -08:00
Daniel Han
15da0659ab
Merge branch 'main' into nightly
2024-11-06 19:02:44 -08:00
Daniel Han
0c8c5ed81e
Update _utils.py
2024-11-06 19:00:42 -08:00
Daniel Han
9a74776712
Update loader.py
2024-11-06 19:00:23 -08:00
Daniel Han
d920393690
Update loader.py
2024-11-06 19:00:13 -08:00
Daniel Han
68b8d623b0
Merge branch 'main' into nightly
2024-11-06 17:21:21 -08:00
Daniel Han
8d6d78fee6
Bug fixes ( #1259 )
...
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
* Update _utils.py
* fix/transformers-unpack (#1180 )
* Fix DPO, ORPO (#1177 )
* Fix TRL
* Update mistral.py
* Patch processing_class
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Update tokenizer_utils.py
* Installation guide (#1165 )
* chore: update chat_templates.py (#1166 )
orginal -> original
* Disable Flex Attention
* Update tokenizer_utils.py
* Update _utils.py
* n_items
* Update cross_entropy_loss.py
* Fix DPO, ORPO
* Update _utils.py
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Add warning for missing Unpack and KwargsForCausalLM in older Transformers versions
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
* Update cross_entropy_loss.py
* Update _utils.py
* Update _utils.py
* donot upcast lm_head and embeddings to float32 (#1186 )
* Cleanup upcast logs (#1188 )
* Fix/phi-longrope (#1193 )
* Enhance rotary embedding handling in LlamaAttention and LongRopeRotaryEmbedding
* Typo
* Improve rotary embedding handling in LlamaAttention to prevent errors with short KV cache
* Update llama.py
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Update transformers
* Unk token issues
* Update _utils.py
* Fix pad token
* Update llama.py
* Typo
* ignored labels
* Revert "ignored labels"
This reverts commit 9d07be077b#1219 )
* Update save.py
Check whether path is in /tmp dir for Kaggle environment
* Update save.py
Move temporary_location to /tmp in Kaggle
* Enhance Kaggle environment support in save and tokenizer utilities
---------
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
* Bug fixes
* Update pyproject.toml
* Update _utils.py
* Update __init__.py
* Update __init__.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update _utils.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Update cross_entropy_loss.py
* Tied weights
* Revert "Tied weights"
This reverts commit 8090b7c01a#1254 )
* Fix: cast logits to float32 in cross_entropy_forward to prevent errors
* Update cross_entropy_loss.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* Throw error when inferencing longer than max_popsition_embeddings (#1236 )
* Throw error when inferencing longer than max_popsition_embeddings without rope scaling
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
* CLI now handles user input strings for dtype correctly (#1235 )
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
* Update flex_attention.py
* Update _utils.py
* Update _utils.py
* Update flex_attention.py
* Update flex_attention.py
* Update loader.py
* Update loader.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update flex_attention.py
* Update _utils.py
---------
Co-authored-by: timothelaborie <97834767+timothelaborie@users.noreply.github.com>
Co-authored-by: Ikko Eltociear Ashimine <eltociear@gmail.com>
Co-authored-by: Edd <68678137+Erland366@users.noreply.github.com>
Co-authored-by: Datta Nimmaturi <datta.nimmaturi@nutanix.com>
Co-authored-by: dendarrion <37800703+dendarrion@users.noreply.github.com>
Co-authored-by: Erland366 <erland.pg366@gmail.com>
Co-authored-by: Edwin Fennell <edwinfennell1@gmail.com>
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
2024-11-06 17:17:19 -08:00
Daniel Han
52bca32ce9
Update _utils.py
2024-11-06 17:14:56 -08:00
Daniel Han
ef5052a8bc
Update flex_attention.py
2024-11-06 16:59:15 -08:00
Daniel Han
4b01ff1724
Update flex_attention.py
2024-11-06 15:56:52 -08:00
Daniel Han
ce621b7af3
Update flex_attention.py
2024-11-06 15:56:26 -08:00
Daniel Han
501ca84297
Update flex_attention.py
2024-11-06 15:39:32 -08:00
Daniel Han
3316ee2282
Update loader.py
2024-11-06 15:13:03 -08:00
Daniel Han
da61c4dde9
Update loader.py
2024-11-06 15:05:10 -08:00
Daniel Han
1ba9f2ed87
Update flex_attention.py
2024-11-06 14:54:53 -08:00
Daniel Han
6b5c599394
Update flex_attention.py
2024-11-06 14:51:52 -08:00
Daniel Han
7e0877d383
Update _utils.py
2024-11-06 14:49:37 -08:00
Daniel Han
05170cd81a
Update _utils.py
2024-11-06 14:05:52 -08:00
Daniel Han
be72975d8a
Update flex_attention.py
2024-11-06 13:09:29 -08:00
Edwin Fennell
de1049bcc5
CLI now handles user input strings for dtype correctly ( #1235 )
...
Co-authored-by: root <root@ieeres.chu.cam.ac.uk>
2024-11-06 12:23:09 -08:00
Datta Nimmaturi
3d906e6378
Throw error when inferencing longer than max_popsition_embeddings ( #1236 )
...
* Throw error when inferencing longer than max_popsition_embeddings without rope scaling
* Update llama.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-06 12:22:08 -08:00
Edd
f24aef5cbc
Fix: cast logits to float32 in cross_entropy_forward to prevent errors ( #1254 )
...
* Fix: cast logits to float32 in cross_entropy_forward to prevent errors
* Update cross_entropy_loss.py
---------
Co-authored-by: Daniel Han <danielhanchen@gmail.com>
2024-11-06 12:16:02 -08:00
