DCAgent/e1_random_d1_original_sandboxes
DCAgent/e1_random_d1_original_sandboxes is an 8 billion parameter language model fine-tuned from Qwen/Qwen3-8B. This model was trained on a specific dataset derived from 'e1_random_d1_original_sandboxes_glm_4.7_traces_jupiter', suggesting a specialization in processing or generating content related to its training data. It is designed for tasks aligned with the characteristics of its fine-tuning dataset, offering capabilities inherited from the Qwen3-8B architecture.
Loading preview...
Model Overview
DCAgent/e1_random_d1_original_sandboxes is an 8 billion parameter language model, fine-tuned from the base Qwen/Qwen3-8B architecture. This model has undergone specialized training on the /e/scratch/jureap59/raoof1/sft_data/hf_hub/datasets--DCAgent--e1_random_d1_original_sandboxes_glm_4.7_traces_jupiter/snapshots/b3b4749051c82e27744d13bfc5e3cd416d8d46f8_thinking_preprocessed dataset.
Training Details
The fine-tuning process utilized the following key hyperparameters:
- Learning Rate: 4e-05
- Optimizer: ADAMW_TORCH_FUSED with betas=(0.9, 0.98) and epsilon=1e-08
- Batch Size: 1 (train), 8 (eval) per device, totaling 16 (train) and 128 (eval) across 16 GPUs
- Epochs: 7.0
- Scheduler: Cosine learning rate scheduler with a 0.1 warmup ratio
This model is built upon Transformers 4.57.6, Pytorch 2.9.1+cu130, Datasets 4.7.0, and Tokenizers 0.22.2. Further details regarding its specific intended uses, limitations, and comprehensive evaluation data are not provided in the current documentation.