train_stsb_789_1760637927

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

• Loss: 4.6980
• Num Input Tokens Seen: 8752512

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: 4
• eval_batch_size: 4
• seed: 789
• 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
4.7781	1.0	1294	4.8901	437440
4.8143	2.0	2588	4.7481	875776
4.7144	3.0	3882	4.7171	1313392
4.5012	4.0	5176	4.7069	1753440
4.6165	5.0	6470	4.7115	2191488
4.81	6.0	7764	4.7011	2629728
4.4637	7.0	9058	4.7121	3066368
4.5863	8.0	10352	4.7064	3502736
4.8275	9.0	11646	4.7091	3940080
4.7608	10.0	12940	4.7056	4376768
4.735	11.0	14234	4.6980	4813216
4.6056	12.0	15528	4.7027	5251344
4.8809	13.0	16822	4.7084	5689568
4.7893	14.0	18116	4.7174	6127200
4.5069	15.0	19410	4.7074	6564352
4.6891	16.0	20704	4.7177	7002400
4.7261	17.0	21998	4.7219	7438560
4.7895	18.0	23292	4.7196	7876400
4.7149	19.0	24586	4.7196	8314032
4.7564	20.0	25880	4.7196	8752512

Framework versions

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