train_copa_123_1760637645

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: 0.2329
• Num Input Tokens Seen: 563328

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: 0.001
• train_batch_size: 4
• eval_batch_size: 4
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.2863	1.0	90	0.2360	28096
0.2354	2.0	180	0.2333	56128
0.2629	3.0	270	0.2331	84352
0.2408	4.0	360	0.2304	112576
0.2297	5.0	450	0.2377	140832
0.232	6.0	540	0.2334	169056
0.2277	7.0	630	0.2333	197344
0.2353	8.0	720	0.2349	225536
0.2399	9.0	810	0.2331	253696
0.2439	10.0	900	0.2344	281856
0.2264	11.0	990	0.2322	310080
0.2338	12.0	1080	0.2341	338144
0.2345	13.0	1170	0.2311	366336
0.2287	14.0	1260	0.2354	394464
0.2284	15.0	1350	0.2347	422592
0.2262	16.0	1440	0.2310	450624
0.2316	17.0	1530	0.2371	478720
0.2338	18.0	1620	0.2334	507008
0.2265	19.0	1710	0.2336	535136
0.2275	20.0	1800	0.2341	563328

Framework versions

• PEFT 0.17.1
• Transformers 4.51.3
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4