Tonic/SmolFactory at c23e2f561499cd8aa4fd39aaa159470df70cbdc5

SmolFactory / config

115 kB

3 contributors

adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing

59e57ff 3 months ago

__init__.py

1.8 kB

adds max perf script 3 months ago
runpod_config.py

1.27 kB

adds no think tag correctly 3 months ago
train_gpt_oss_basic.py

5.67 kB

adds gpt-oss support 3 months ago
train_gpt_oss_custom.py

18.7 kB

adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 3 months ago
train_gpt_oss_h100_optimized.py

7.45 kB

adds gpt-oss support 3 months ago
train_gpt_oss_memory_optimized.py

5.32 kB

adds quantization configuration correctly 3 months ago
train_gpt_oss_multilingual_reasoning.py

7.84 kB

adds gpt-oss support 3 months ago
train_gpt_oss_openhermes_fr.py

8.13 kB

adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 3 months ago
train_gpt_oss_openhermes_fr_memory_optimized.py

12.3 kB

adds harmony format , configurable gpt-oss parameters, launch.sh logic , improved templates for legml gpt-oss training, dynamic results directory and improve model pushing 3 months ago
train_smollm3.py

3.64 kB

adds sft , quantization, better readmes 3 months ago
train_smollm3_dpo.py

3.38 kB

adds sft , quantization, better readmes 3 months ago
train_smollm3_h100_lightweight.py

6.47 kB

attempt to fix bfloat16 issue 3 months ago
train_smollm3_long_context.py

1.02 kB

adds no think tag correctly 3 months ago
train_smollm3_no_think_test.py

914 Bytes

adds no think tag correctly 3 months ago
train_smollm3_openhermes_fr.py

4.48 kB

adds formatting fix 3 months ago
train_smollm3_openhermes_fr_a100_balanced.py

7 kB

adds formatting fix 3 months ago
train_smollm3_openhermes_fr_a100_large.py

6.51 kB

adds formatting fix 3 months ago
train_smollm3_openhermes_fr_a100_max_performance.py

6.78 kB

adds formatting fix 3 months ago
train_smollm3_openhermes_fr_a100_multiple_passes.py

6.7 kB

adds formatting fix 3 months ago