train_multirc_1756735871

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the multirc dataset. It achieves the following results on the evaluation set:

• Loss: 0.3394
• Num Input Tokens Seen: 117044976

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 123
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.3178	0.5000	6130	0.3802	5867968
0.5557	1.0001	12260	0.3493	11717296
0.0109	1.5001	18390	0.1670	17588368
0.2324	2.0002	24520	0.3016	23419920
0.277	2.5002	30650	0.3568	29254768
0.3393	3.0002	36780	0.3402	35127456
0.3145	3.5003	42910	0.3391	40997712
0.3363	4.0003	49040	0.3400	46843184
0.3059	4.5004	55170	0.3385	52692880
0.4232	5.0004	61300	0.3394	58550736
0.2604	5.5004	67430	0.3361	64403536
0.2416	6.0005	73560	0.3536	70253968
0.4941	6.5005	79690	0.3594	76098208
0.4356	7.0006	85820	0.3684	81942784
0.5113	7.5006	91950	0.3346	87796256
0.2519	8.0007	98080	0.3266	93652576
0.388	8.5007	104210	0.3368	99519344
0.4701	9.0007	110340	0.3482	105351168
0.4091	9.5008	116470	0.3413	111213008

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1