train_copa_1756729609

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

• Loss: 0.2389
• Num Input Tokens Seen: 273712

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
0.2146	0.5	90	0.2563	13664
0.2358	1.0	180	0.2523	27408
0.2177	1.5	270	0.2437	41120
0.2349	2.0	360	0.2347	54752
0.2026	2.5	450	0.2439	68432
0.2456	3.0	540	0.2322	82176
0.2229	3.5	630	0.2402	95936
0.2258	4.0	720	0.2348	109584
0.2307	4.5	810	0.2455	123232
0.2319	5.0	900	0.2316	137008
0.2225	5.5	990	0.2376	150672
0.2297	6.0	1080	0.2333	164336
0.2299	6.5	1170	0.2325	178032
0.2122	7.0	1260	0.2332	191712
0.2274	7.5	1350	0.2341	205312
0.2397	8.0	1440	0.2398	219072
0.2326	8.5	1530	0.2392	232768
0.2314	9.0	1620	0.2372	246416
0.2125	9.5	1710	0.2374	260112
0.2223	10.0	1800	0.2389	273712

Framework versions

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