HYU-NLP-EVAL/20261016_rar-medicine_static_qwen3-4b_norm3200_impl-verl_s11_step-068

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 17, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The HYU-NLP-EVAL/20261016_rar-medicine_static_qwen3-4b_norm3200_impl-verl_s11_step-068 is a 4 billion parameter model based on the Qwen3 architecture, developed by HYU-NLP-EVAL. This model is a specific checkpoint (step 68, seed 11) from the '261016_medicine_static_pilot' run, featuring a context length of 32768 tokens. It is provided as a BF16 Transformers export, with original native veRL/FSDP checkpoint files also available, suggesting a focus on research or specialized applications in the medical domain.

Loading preview...

Overview

This model, HYU-NLP-EVAL/20261016_rar-medicine_static_qwen3-4b_norm3200_impl-verl_s11_step-068, is a 4 billion parameter variant of the Qwen3 architecture. It represents checkpoint step 68 from a specific training run identified as 261016_medicine_static_pilot, utilizing seed 11. The model is provided in a BF16 Transformers export format, making it readily usable within the Hugging Face ecosystem. The original native veRL/FSDP checkpoint files and associated metadata are also included for detailed inspection or further development.

Key Capabilities

  • Specialized Checkpoint: This is a specific snapshot (step 68) from a pilot run, indicating a focused development phase.
  • Qwen3 Architecture: Leverages the capabilities of the Qwen3 model family.
  • BF16 Export: Provided in a common format for easy integration and deployment.
  • Original Checkpoint Access: Includes native veRL/FSDP files for advanced users or research purposes.

Good for

  • Research and Development: Ideal for researchers exploring specific training stages or fine-tuning experiments within the medical domain.
  • Reproducibility: Provides exact checkpoint files for replicating specific training outcomes.
  • Specialized Applications: Potentially suitable for tasks related to medicine, given the pilot run's naming convention, though specific performance metrics are not detailed.

Finetunes
1 models