Model Overview

This model is a fine-tuned iteration of the Qwen3-8B base model, originally developed by Qwen. The fine-tuning process utilized the /e/data1/datasets/playground/ot/hf_hub/datasets--penfever--Kimi-2.5-inferredbugs-sandboxes-maxeps-32k/snapshots/b8446f4e9f3c6a0a77d8866517b15b4ddeb7647d_thinking_preprocessed dataset.

Training Details

The fine-tuning procedure involved specific hyperparameters:

• Learning Rate: 4e-05
• Batch Sizes: train_batch_size of 1, eval_batch_size of 8
• Gradient Accumulation: 3 steps, leading to a total_train_batch_size of 96
• Optimizer: ADAMW_TORCH_FUSED with betas=(0.9, 0.98) and epsilon=1e-08
• LR Scheduler: Cosine type with a warmup ratio of 0.1
• Epochs: 7.0
• Distributed Training: Multi-GPU setup across 32 devices.

Current Status

Further information regarding the model's specific description, intended uses, limitations, and detailed training/evaluation data is currently pending. Users should consult future updates for a comprehensive understanding of its capabilities and optimal application scenarios.