LSW142857/OPSD-PI-Qwen3.5-9B-Medium-Trailing-1024-A6000-Merged-Update08

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 2, 2026Architecture:Transformer Featherless Exclusive Cold

LSW142857/OPSD-PI-Qwen3.5-9B-Medium-Trailing-1024-A6000-Merged-Update08 is a 9 billion parameter Qwen3.5-based causal language model developed by LSW142857. This model represents the eighth optimizer update from an OPSD training run, incorporating merged expert-SFT initialization and LoRA updates. It is designed for direct loading without additional merging steps, making it suitable for applications requiring a pre-integrated model from a specific training iteration.

Loading preview...

Model Overview

LSW142857/OPSD-PI-Qwen3.5-9B-Medium-Trailing-1024-A6000-Merged-Update08 is a 9 billion parameter model derived from the Qwen3.5 architecture. This specific release is the result of 8 completed optimizer updates (training iteration 7) from a 1024-row Medium PI trailing_user OPSD run conducted on 8×RTX A6000 GPUs. The model is provided as a fully merged Hugging Face repository, meaning it can be loaded directly without requiring any adapter or additional merge steps.

Key Characteristics

  • Directly Loadable: The repository contains four merged model shards, integrating expert-SFT initialization, OPSD main-model LoRA updates, MTP LoRA updates, and directly trained full-MTP tensors. This simplifies deployment as no further merging is needed.
  • Training Provenance: The model's integrity and provenance are verifiable through sha256sum checks and detailed merge_manifest.json and training_config.json files, which document hashes, source identity, configuration, and finite metrics from this specific update.
  • Teacher-Only PI: During its training, the PI (Private Information) was teacher-only. Users are advised to evaluate the student model without adding PI and to use held-out tasks rather than the 1024 training rows for assessment.

Usage Considerations

This model is suitable for developers who need a specific snapshot of a Qwen3.5-based model after a defined number of OPSD training updates. Its pre-merged state streamlines integration into existing workflows. For evaluation, it is crucial to use datasets distinct from the training data to ensure unbiased performance assessment.

Finetunes
1 models