ROCm / ROCm/FastFlowLM

Feature Request: Add Mistral AI Model Family Support (Mistral 7B, Codestral, Voxtral Mini)

Open
#340 3 comments 6 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

model request
Dominant language
C++
Stars
1.9k
Forks
152
Avg merge
4h 14m
Merged PRs (30d)
11

Description

Summary

Request for Mistral AI model support in FastFlowLM. AMD has already published NPU-optimized ONNX versions of Mistral 7B, reducing integration effort.

Requested Models

Model Size Use Case NPU-Ready
Mistral 7B Instruct v0.3 7B General chat/instruction AMD ONNX available
Codestral 25.08 22B Code generation, FIM completion ❌ Needs optimization
Voxtral Mini 3B Speech-to-text, audio Q&A ❌ Needs optimization

Why Mistral?

  • Low effort for Mistral 7B: AMD's NPU-optimized ONNX already exists (AWQ INT4, BF16)
  • Fills gaps in current model lineup:
    • Codestral → first code-specialized model in FLM
    • Voxtral Mini → complements Whisper with native Q&A/summarization
  • Licensing: Voxtral Mini is Apache 2.0; Mistral 7B permissive for local use

References

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are named. Start by reviewing how existing model families are integrated and how AMD NPU-optimized ONNX models are loaded, then determine the separate requirements for Mistral 7B, Codestral, and Voxtral Mini; done means supported model paths and validation for all requested families.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp, huggingface
Domain
ai-infra-agents, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.