FRPO/qwen3-1.7b-a16_onpolicy_seqmean-k1-cNone-clip0.2-mb1-eta100-bs64x5-n2

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 15, 2026Architecture:Transformer Featherless Exclusive Cold

FRPO/qwen3-1.7b-a16_onpolicy_seqmean-k1-cNone-clip0.2-mb1-eta100-bs64x5-n2 is a 2 billion parameter language model based on the Qwen3-1.7B architecture, fine-tuned using Reinforcement Learning (RL) with the FRPO experimental framework. This model is a specific checkpoint from the KL-in-LLM-RL experiments, utilizing the verl training library. It is designed for research and development in RL-based language model optimization, offering a specific configuration for on-policy sequence mean training.

Loading preview...

Model Overview

This model, FRPO/qwen3-1.7b-a16_onpolicy_seqmean-k1-cNone-clip0.2-mb1-eta100-bs64x5-n2, is an RL fine-tuned checkpoint derived from the KL-in-LLM-RL / FRPO experimental series. It is built upon the Qwen/Qwen3-1.7B base model, featuring approximately 2 billion parameters and a context length of 32768 tokens. The fine-tuning process was conducted using the verl library.

Key Characteristics

  • Base Model: Qwen3-1.7B, a robust foundation for language understanding and generation.
  • Fine-tuning Method: Utilizes Reinforcement Learning (RL) within the FRPO (Fictitious Reinforcement Policy Optimization) framework.
  • Specific Checkpoint: Represents global_step_800 from the training run, provided in fp32 safetensors format without post-processing.
  • Configuration: The model's specific training configuration, including parameters like a16_onpolicy_seqmean-k1-cNone-clip0.2-mb1-eta100-bs64x5-n2, is encoded directly in its repository name, indicating its experimental nature.

Intended Use Cases

This model is primarily intended for:

  • RL Research: Researchers exploring Reinforcement Learning techniques for language model optimization, particularly within the KL-in-LLM-RL and FRPO contexts.
  • Experimental Analysis: Developers and researchers interested in analyzing the effects of specific RL configurations (e.g., on-policy sequence mean, clip values, batch sizes) on language model performance.
  • Comparative Studies: As a specific checkpoint, it can be used to compare different RL fine-tuning strategies or hyperparameter settings within the verl framework.

Finetunes
1 models