lc-7b-sft-lora-full

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 1.5104

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
1.5657	1.0	31	1.6190
1.5249	2.0	62	1.5615
1.4901	3.0	93	1.5294
1.4397	4.0	124	1.5111
1.3942	5.0	155	1.5020
1.3161	6.0	186	1.4975
1.3081	7.0	217	1.4987
1.352	8.0	248	1.4962
1.3162	9.0	279	1.4946
1.3019	10.0	310	1.5006
1.296	11.0	341	1.5009
1.2174	12.0	372	1.5040
1.3063	13.0	403	1.5075
1.294	14.0	434	1.5082
1.2651	15.0	465	1.5086
1.2766	16.0	496	1.5095
1.24	17.0	527	1.5098
1.2455	18.0	558	1.5100
1.33	19.0	589	1.5099
1.2447	20.0	620	1.5104

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.1.0+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1