ProsperoAI-Mistral-7B-Instruct-v0.3-Q4_0-PS5

ProsperoAI Mistral 7B Instruct v0.3 Q4_0 for PS5

This repository contains a curated, install-ready conversion of Mistral 7B Instruct v0.3 Q4_0 for the ProsperoAI native PlayStation 5 GPU runtime. It is intended for local inference with the ProsperoAI homebrew app and does not use a cloud API.

This is a runtime-format conversion, not a fine-tune. The model weights come from the Q4_0 GGUF published by QuantFactory, based on mistralai/Mistral-7B-Instruct-v0.3. Both sources are distributed under the Apache 2.0 license.

Target runtime

• Application: ProsperoAI (PPSA99004)
• Processor: PlayStation 5 AMD GPU through the native AGC compute path
• Runtime profile: prosperoai-mistral-7b-v1
• Container format: P5LM v1, runtime layout mistral-7b-runtime-v1
• Quantization: GGUF Q4_0 weights, with the source model's F32 and Q6_K tensors
• Execution: GPU-resident weights and KV cache; CPU orchestration and tokenization
• Current app context limit: 4,096 tokens
• Current decoding: greedy token selection

This P5LM file is not a GGUF file and is not directly compatible with llama.cpp, ROCm, CUDA, or desktop model runners. It is laid out specifically for ProsperoAI's native PS5 AGC kernels.

Validated architecture

The converter and runtime validate this exact profile before accepting a model:

• 32 transformer layers
• embedding width 4,096
• feed-forward width 14,336
• 32 attention heads and 8 KV heads
• vocabulary size 32,768
• RoPE dimension 128 and frequency base 1,000,000
• 291 tensors in the expected Mistral 7B layout

Install

Download the repository, then copy the complete folder mistral-7b-instruct-v0.3-q4-0 into the app's models directory:

PPSA99004/
└── models/
    └── mistral-7b-instruct-v0.3-q4-0/
        ├── model.ps5lm
        ├── tokenizer.ps5tok
        └── model.json

Keep the generic filenames unchanged. The named parent folder identifies the model and allows ProsperoAI to discover multiple installed models. Restart the app after copying or replacing a model.

Files and integrity

Source GGUF:

• Repository: QuantFactory/Mistral-7B-Instruct-v0.3-GGUF
• File: Mistral-7B-Instruct-v0.3.Q4_0.gguf
• Revision: a83fbb7c79c1afd59dfcce75272591cbe5a2412f
• SHA-256: 19c3cf872246d48e7a9784d6def7a694a6c9056cfbeb98f7061d43cd0a7c8b7f

The full machine-readable conversion record is retained separately at provenance/model-preparation.json; it is not required on the PS5.

Reproduce the conversion

Download the exact source GGUF listed above and the ProsperoAI Alpha 2 model tools, then run:

prepare-model --provenance provenance/model-preparation.json \
  Mistral-7B-Instruct-v0.3.Q4_0.gguf

The recipe contains the complete validated tensor layout. The tool verifies the source and all three generated files against their recorded SHA-256 values.

Performance notes

Short-context generation is substantially faster than long-context generation. As the conversation grows, attention work per generated token increases and new prompt suffixes must be prefetched into the KV cache. ProsperoAI reuses the matching cached prefix, but throughput still declines with context length. The 4,096-token limit is a functional ceiling, not a promise of constant speed at all context sizes.

Limitations

• Requires a PS5 environment capable of running the ProsperoAI native app.
• Only the exact architecture and tensor profile above is currently supported.
• The app's context limit is lower than the upstream model's advertised limit.
• Q4_0 quantization trades some model quality for memory use and speed.
• Follow the upstream Mistral model card, license, and acceptable-use guidance.

Curated and converted for ProsperoAI by BlackBearReloaded.