train_mnli_42_1773765555

This model is a fine-tuned version of meta-llama/Llama-3.2-1B-Instruct on the mnli dataset. It achieves the following results on the evaluation set:

• Loss: 0.2161
• Num Input Tokens Seen: 191491960

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.259	0.2500	11045	0.2206	9565376
0.2626	0.5000	22090	0.2597	19168640
0.3184	0.7500	33135	0.2563	28714304
0.237	1.0000	44180	0.2458	38289824
0.2216	1.2500	55225	0.2485	47877216
0.1684	1.5000	66270	0.2347	57416032
0.1456	1.7500	77315	0.2226	66982176
0.1218	2.0000	88360	0.2205	76602496
0.2176	2.2501	99405	0.2270	86154496
0.1293	2.5001	110450	0.2184	95709312
0.1849	2.7501	121495	0.2161	105304960
0.1495	3.0001	132540	0.2170	114898176
0.1305	3.2501	143585	0.2544	124468928
0.0799	3.5001	154630	0.2460	134028992
0.1014	3.7501	165675	0.2478	143607232
0.1224	4.0001	176720	0.2528	153206432
0.1156	4.2501	187765	0.3163	162770528
0.0399	4.5001	198810	0.3196	172345120
0.121	4.7501	209855	0.3213	181948192

Framework versions

• Transformers 4.51.3
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4