sam2ai/DeepSeek-R1-Distill-Qwen-7B-Multilingual-indic

reasoning-multilingual-R1-Llama-70B-train

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-7B on the reasoning-multilingual-R1-Llama-70B-train dataset. It achieves the following results on the evaluation set:

• Loss: 0.5135

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 8
• total_eval_batch_size: 8
• 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: 0.01
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.5539	0.3158	6	0.3874
0.5069	0.6316	12	0.4255
0.5634	0.9474	18	0.4700
0.7207	1.2632	24	0.5061
0.6525	1.5789	30	0.5270
0.7435	1.8947	36	0.5325
0.7488	2.2105	42	0.5383
0.5786	2.5263	48	0.5181
0.7325	2.8421	54	0.5146

Framework versions

• Transformers 4.46.1
• Pytorch 2.6.0.dev20241113+rocm6.2
• Datasets 3.1.0
• Tokenizers 0.20.3