1.26 GB

Ctrl+K

1 contributor

History: 34 commits

Shawnno

v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1%

11fc04f verified 3 months ago

checkpoint-400
v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1% 3 months ago
checkpoint-546
v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1% 3 months ago
checkpoint-578
SFT v2: Qwen2.5-3B QLoRA, eval_loss=0.983, acc=73.9% 3 months ago
.gitattributes
1.57 kB
Upload folder using huggingface_hub 3 months ago
README.md
5.1 kB
SFT v2: Qwen2.5-3B QLoRA, eval_loss=0.983, acc=73.9% 3 months ago
adapter_config.json
798 Bytes
v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1% 3 months ago
adapter_model.safetensors
120 MB
xet

v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1% 3 months ago
added_tokens.json
605 Bytes
v2: sft quality improvements - loss masking, neftune, system prompt, oversampling 3 months ago
merges.txt
1.67 MB
v2: sft quality improvements - loss masking, neftune, system prompt, oversampling 3 months ago
special_tokens_map.json
496 Bytes
v2: sft quality improvements - loss masking, neftune, system prompt, oversampling 3 months ago
tokenizer.json
11.4 MB
xet

SFT v2: Qwen2.5-3B QLoRA, eval_loss=0.983, acc=73.9% 3 months ago
tokenizer_config.json
7.33 kB
SFT v2: Qwen2.5-3B QLoRA, eval_loss=0.983, acc=73.9% 3 months ago
train.log
60.9 kB
v3 SFT: lr=2e-4 NEFTune=0 no-oversample | eval_loss 1.039->0.992 train_loss 0.941 acc 81.1% 3 months ago
vocab.json
2.78 MB
v2: sft quality improvements - loss masking, neftune, system prompt, oversampling 3 months ago