add model card

README.md

@@ -0,0 +1,68 @@
+---
+license: mit
+base_model: MiniMaxAI/MiniMax-M2.7
+tags:
+- hipfire
+- moe
+- minimax
+- quantized
+library_name: hipfire
+---
+
+# hipfire-MiniMax-M2.7
+
+[hipfire](https://github.com/Kaden-Schutt/hipfire)-native quantizations of
+[MiniMax-M2.7](https://huggingface.co/MiniMaxAI/MiniMax-M2.7) — a 229B-parameter
+(≈10B active) Mixtral-style MoE: GQA attention with per-layer QK-norm, partial
+rotate_half RoPE, and a 256-expert top-8 sigmoid+bias router (DeepSeek-V3-style),
+SwiGLU experts, no shared expert. 62 layers, hidden 3072, vocab 200064.
+
+These `.mq*` files run with the hipfire inference engine (HIP/ROCm-direct, no
+Python in the hot path). They are **not** GGUF/safetensors and won't load in
+llama.cpp / transformers.
+
+## Files
+
+| File | Expert format | Size | Target |
+|------|---------------|------|--------|
+| `MiniMax-M2.7.mq4` | MQ4G256 (4-bit FWHT) experts, Q8 attn/router/head | 123.6 GB | high quality |
+| `MiniMax-M2.7.mq2-lloyd` | MQ2G256-Lloyd (2-bit + codebook) experts, Q8 attn/router/head | ≈72 GB | fits 128 GB unified (e.g. Strix Halo / gfx1151) |
+
+## Validation
+
+The hipfire MiniMax-M2 forward pass was validated against a PyTorch
+(`transformers`) reference on a dimension-faithful tiny random-weight oracle
+(per-layer post-residual hidden states):
+
+- **mq4 experts**: cosine **0.9996** vs reference (residual = 4-bit quant noise)
+- **mq2-lloyd experts**: cosine **0.9987** vs reference (residual = 2-bit-Lloyd noise)
+
+Attention isolated to cosine 0.99990; routing (sigmoid + bias top-8) matches the
+reference selection and weights exactly.
+
+The real 229B `mq4` build was confirmed end-to-end: it loads (123.6 GB) and
+generates coherent, factually-correct text (e.g. completes "The capital of
+France is" → "Paris. The capital of Germany is Berlin. ...") at ~72 tok/s decode
+on a single MI300X.
+
+## Quantization
+
+Source: the official FP8 (E4M3 + F32 block-[128,128] `weight_scale_inv`)
+checkpoint, dequantized to F32 then re-quantized. Attention projections, router,
+embedding, and lm_head are kept at Q8; routed experts are MQ4G256 (mq4) or
+MQ2G256-Lloyd (mq2-lloyd).
+
+## Usage (hipfire)
+
+```
+hipfire serve --model MiniMax-M2.7.mq2-lloyd
+# or a single forward:
+# examples/infer_minimax --model MiniMax-M2.7.mq4 --prompt "..."
+```
+
+## Notes
+
+- MQ3 / MQ2 / MQ3-Lloyd expert variants are not yet published: hipfire does not
+  currently ship indexed-decode MoE GEMV kernels for those formats (only
+  MQ4G256, MQ6G256, and MQ2G256-Lloyd have complete decode+prefill MoE kernels).
+- arch_id 10 in the hipfire HFQ header.