train_record_42_1773765559

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

• Loss: 0.8647
• Num Input Tokens Seen: 245808128

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
1.2402	0.2500	3906	1.0148	12292032
1.4262	0.5001	7812	1.2947	24620672
1.1304	0.7501	11718	1.2590	36894016
0.9504	1.0002	15624	1.1777	49176512
0.9611	1.2502	19530	1.1499	61465280
0.7862	1.5003	23436	1.0926	73739776
0.8905	1.7503	27342	1.0198	86015936
0.7648	2.0004	31248	0.9767	98341056
0.7543	2.2504	35154	1.0004	110649216
0.5464	2.5005	39060	0.9313	122910592
0.5669	2.7505	42966	0.9105	135222656
0.2979	3.0006	46872	0.8803	147516736
0.21	3.2506	50778	0.9521	159826368
0.4287	3.5007	54684	0.9180	172084032
0.2835	3.7507	58590	0.8755	184402752
0.2202	4.0008	62496	0.8647	196687936
0.1392	4.2508	66402	1.0180	209017024
0.2328	4.5009	70308	0.9995	221278272
0.1733	4.7509	74214	0.9999	233564288

Framework versions

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