JamieBradfield/qwen3.8-9b-hermes-function-calling-v2-GGUF overview
Qwen3.8 9B Hermes FC v2 — GGUF ROCmFPX ROCmFPX quantized GGUF of the BF16 merge https://huggingface.co/JamieBradfield/qwen3.8 9b hermes function calling v2 . F…
Runs locally from ~4.58 GB disk (8 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| qwen3.8-9b-hf-fc-v2-Q4_0_ROCMFP4_FAST.gguf | GGUF | Q4_0_ROCMFP4_FAST | 4.58 GB | Download |
Model Details
| Model ID | JamieBradfield/qwen3.8-9b-hermes-function-calling-v2-GGUF |
|---|---|
| Author | JamieBradfield |
| Pipeline | text-generation |
| License | apache-2.0 |
| Base model | JamieBradfield/qwen3.8-9b-hermes-function-calling-v2 |
| Last modified | 2026-08-28T22:31:33.000Z |
Model README
---
license: apache-2.0
base_model: JamieBradfield/qwen3.8-9b-hermes-function-calling-v2
tags:
- qwen3.5
- gguf
- rocmfpx
- function-calling
pipeline_tag: text-generation
---
Qwen3.8-9B Hermes FC v2 — GGUF (ROCmFPX)
ROCmFPX quantized GGUF of the BF16 merge.
Files
| file | quant | size |
|---|---|---|
| qwen3.8-9b-hf-fc-v2-Q4_0_ROCMFP4_FAST.gguf | Q4_0_ROCMFP4_FAST | 4.92 GB |
Notes
- ROCmFPX quants use ROCMFP4 kernels from the
llama-rocmfpxfork
(AMD RDNA3 / RX 7700 XT). Not portable to CUDA or CPU-only builds without
the fork's kernels.
- MTP (multi-token prediction) head is preserved in the GGUF: the BF16
merge drops only the vision tower.
- Converted from the BF16 merge with
convert_hf_to_gguf.py --outtype bf16,
quantized with llama-quantize Q4_0_ROCMFP4_FAST.
- Model card and evaluation status: see the base-model repo.
Run JamieBradfield/qwen3.8-9b-hermes-function-calling-v2-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models