e29edc5b559dc8630f4e2d1120c68181

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Llama-8B on the nyu-mll/glue [sst2] dataset. It achieves the following results on the evaluation set:

• Loss: 2.9275
• Data Size: 1.0
• Epoch Runtime: 2097.7784
• Accuracy: 0.7882
• F1 Macro: 0.7881
• Rouge1: 0.7876
• Rouge2: 0.0
• Rougel: 0.7870
• Rougelsum: 0.7882

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
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	5.0130	0	3.6490	0.4838	0.4315	0.4826	0.0	0.4838	0.4850
No log	1	2104	3.0521	0.0078	20.3535	0.4907	0.3292	0.4907	0.0	0.4907	0.4919
No log	2	4208	4.4459	0.0156	40.8786	0.5093	0.3374	0.5093	0.0	0.5093	0.5081
0.2038	3	6312	2.8874	0.0312	81.9876	0.6701	0.6532	0.6701	0.0	0.6713	0.6701
2.3299	4	8416	2.6848	0.0625	152.1367	0.6794	0.6519	0.6782	0.0	0.6794	0.6794
1.7086	5	10520	2.0967	0.125	283.0299	0.7523	0.7428	0.7523	0.0	0.7512	0.7523
1.1936	6	12624	1.7836	0.25	529.6649	0.8113	0.8113	0.8113	0.0	0.8113	0.8113
1.0825	7	14728	2.0926	0.5	1048.1050	0.7951	0.7949	0.7946	0.0	0.7951	0.7951
0.8172	8.0	16832	1.9382	1.0	2090.7046	0.7986	0.7978	0.7975	0.0	0.7986	0.7986
0.5088	9.0	18936	2.9781	1.0	2087.0249	0.7951	0.7948	0.7940	0.0	0.7951	0.7951
0.427	10.0	21040	2.9275	1.0	2097.7784	0.7882	0.7881	0.7876	0.0	0.7870	0.7882

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1