Tonic/SmolFactory at 598357a7d8a087543477f0aac96b9910dc8dbbef

1.1 MB

3 contributors

improves launcher with model family and defaults based on options, updates trl trainer , removes trl config paths by switching to trainingarguments class , tokenizer parameters updated to sfttrainer , resolves evaluation_strategy error

598357a 4 months ago

config
adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 4 months ago
docs
adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 4 months ago
requirements
adds quantization configuration correctly 4 months ago
scripts
improves launcher with model family and defaults based on options, updates trl trainer , removes trl config paths by switching to trainingarguments class , tokenizer parameters updated to sfttrainer , resolves evaluation_strategy error 4 months ago
src
adds single token logic read/write , adds gpt-oss demo space , adds spaces refactor , adds new version of track tonic , adds logic in launch.sh 4 months ago
templates
adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 4 months ago
test_data
Fix model recovery and deployment scripts - add safetensors support and Windows compatibility 4 months ago
tests
adds single token logic read/write , adds gpt-oss demo space , adds spaces refactor , adds new version of track tonic , adds logic in launch.sh 4 months ago
.gitignore

903 Bytes

adds more documentation 4 months ago
README.md

11.8 kB

adds gpt-oss support 4 months ago
launch.sh

49.4 kB

improves launcher with model family and defaults based on options, updates trl trainer , removes trl config paths by switching to trainingarguments class , tokenizer parameters updated to sfttrainer , resolves evaluation_strategy error 4 months ago
trackio.py

657 Bytes

revert file rename 4 months ago