foundation-model-stack fms-hf-tuning issues

foundation-model-stack / fms-hf-tuning

🚀 Collection of tuning recipes with HuggingFace SFTTrainer and PyTorch FSDP.

Apache License 2.0

9 stars 30 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

initial deploy script for testing tuning and inference

#222 anhuong opened 3 hours ago
0
Update transformers requirement from !=4.38.2,<=4.40.2,>=4.34.1 to !=4.38.2,<=4.42.2,>=4.42.2

#221 dependabot[bot] opened 20 hours ago
0
build(deps): Update transformers requirement from !=4.38.2,<=4.40.2,>=4.34.1 to !=4.38.2,<=4.42.1,>=4.42.1

#220 dependabot[bot] closed 20 hours ago
1
Fix PyPi publish error caused by direct url reference

#219 tedhtchang closed 20 hours ago
2
deps: cap transformers at 4.40.2

#218 anhuong closed 1 day ago
0
bug: Pod runs out of ephemeral storage (disk) space because of the temporary directory.

#217 HarikrishnanBalagopal opened 2 days ago
0
Formatting consolidation main

#216 Ssukriti closed 1 day ago
0
feat: Need to support aync logging in the library to record metadata and logs from execution.

#215 dushyantbehl opened 3 days ago
0
docs: instructions for using the trainer controller framework

#214 HarikrishnanBalagopal opened 3 days ago
1
Update trl

#213 alex-jw-brooks closed 2 days ago
0
Update packaging requirement from <24,>=23.2 to >=23.2,<25

#212 dependabot[bot] opened 1 week ago
0
Update numpy requirement from <2.0,>=1.26.4 to >=1.26.4,<3.0

#211 dependabot[bot] opened 1 week ago
0
Bump trl from 0.8.6 to 0.9.4

#210 dependabot[bot] closed 2 days ago
1
build: use poetry for reproducible virtual environments

#209 VassilisVassiliadis opened 1 week ago
1
add dependabot.yml

#208 tedhtchang closed 1 week ago
0
Delete dependabot.yml

#207 tedhtchang closed 1 week ago
0
Update trl library past v0.8.6

#206 anhuong closed 2 days ago
1
Improve Acceleration Framework Integration

#205 fabianlim opened 1 week ago
0
DO NOT MERGE: test

#204 anhuong closed 1 week ago
0
deps: pin transformers below v4.41

#203 anhuong closed 1 week ago
1
Disallow installing `transformers >= 4.41`

#202 kpouget closed 1 week ago
7
Performance regression in fms-hf-tuning image

#201 kpouget opened 1 week ago
3
NCCL ENV support

#200 bbenshab opened 1 week ago
0
Fix additional callbacks

#199 VassilisVassiliadis closed 3 days ago
5
Upgrading to TRL>0.9 Causes Incompatibility with Extended TrainingArguments

#198 achew010 opened 2 weeks ago
1
remove merge model for lora tuned adapters

#197 anhuong closed 2 weeks ago
0
fix: freeze to specific set of versions with reliable tighter lower and upper bounds.

#196 kmehant opened 2 weeks ago
2
feat: Allow specifying logging level from the CLI

#195 kmehant opened 2 weeks ago
1
feat: use native python logger instead of transformers logger

#194 kmehant opened 2 weeks ago
0
feat: resizing the embedding layer to the power of 2 or multiple of 8

#193 kmehant opened 2 weeks ago
0
feat: support adding attention mask when not present for pretokenized data

#192 kmehant opened 2 weeks ago
0
feat: support pretokenised datasets

#191 kmehant opened 2 weeks ago
1
feat: support custom tokenizers

#190 kmehant opened 2 weeks ago
0
bug: Pass kwargs to __init__ within derived controls and change operations constructor to accept name and kwargs

#189 seshapad opened 2 weeks ago
0
bug: Added comma after argument in `TrainerControllerCallback`

#188 seshapad opened 2 weeks ago
0
bug: Assert string formats in testcases

#187 seshapad opened 2 weeks ago
0
bug: conditionals should use “is” instead of ==

#186 seshapad opened 2 weeks ago
0
bug: Modify should_() signature in test case to make arguments unnamed

#185 seshapad opened 2 weeks ago
0
bug: Modify should_perform_action_xyz() signature in test case to make arguments unnamed

#184 seshapad opened 2 weeks ago
0
bug: Modify compute() function signature in test case to make arguments unnamed

#183 seshapad opened 2 weeks ago
0
feat: Extend trainer controller capabilities to have fine-grained control in multi-node-multi-gpu scenario

#182 seshapad opened 2 weeks ago
0
feat: Add metric for PerProcessState

#181 seshapad opened 2 weeks ago
0
bug: Test case for “loss not available”

#180 seshapad opened 2 weeks ago
0
feat: Add custom logging operation

#179 seshapad opened 2 weeks ago
0
bug: Missing on_save() event in callback.py

#178 seshapad opened 2 weeks ago
0
feat: Format changes to trainer controller yaml to use ‘_’ instead of ‘-’ in keys

#177 seshapad opened 2 weeks ago
0
feat: Return entire log-line in loss.py metric

#176 seshapad opened 2 weeks ago
0
feat: Pass trainer controller metrics to operations

#175 seshapad opened 2 weeks ago
0
Update README.md for Lora modules

#174 Ssukriti closed 2 weeks ago
0
fix: bloom model can't run with flash-attn

#173 anhuong closed 2 weeks ago
0