mirror of
https://github.com/storytold/cloud-worker.git
synced 2026-10-09 00:09:43 +00:00
Hanashi
68 lines
4.9 KiB
Markdown
68 lines
4.9 KiB
Markdown
# MiniMax H3 — weights log (B200 pod `c306014998a3`, 216.243.220.136)
|
|
|
|
Status as of **2026-08-06 04:35 UTC**.
|
|
|
|
## Source
|
|
|
|
All files downloaded from the HuggingFace repo
|
|
**[Comfy-Org/MiniMax-H3](https://huggingface.co/Comfy-Org/MiniMax-H3)** (the
|
|
ComfyUI repack of [MiniMaxAI/MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3)),
|
|
e.g. `https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/<path>`.
|
|
Downloaded with `hf download` + hf_transfer by `scripts/pod/download-weights.sh`
|
|
and `download-weights-tmpfs.sh`; log at `/workspace/minimax-h3/download.log`.
|
|
|
|
There are two model lines — **fl2va** (text-to-video + first/last-frame i2v)
|
|
and **ref2va** (reference-to-video) — each published in several precisions.
|
|
ComfyUI finds both storage locations below via `/ComfyUI/extra_model_paths.yaml`
|
|
(installed from `scripts/pod/extra_model_paths.yaml`).
|
|
|
|
## Storage locations on the pod
|
|
|
|
| Path | Medium | Notes |
|
|
|---------------------------------|------------------------------|----------------------------------------------|
|
|
| `/workspace/minimax-h3/models/` | RunPod network volume (1 TB) | **Persistent** — survives pod resets/edits. |
|
|
|
|
The volume was resized 100 GB → 1 TB on 2026-08-06 (which reset the pod), so
|
|
the earlier `/dev/shm` tmpfs overflow scheme is retired: **everything now
|
|
lives on the volume.**
|
|
|
|
## Downloaded — all files, `/workspace/minimax-h3/models/` (246 GB total)
|
|
|
|
| File | Family | Size |
|
|
|----------------------------------------------------------------------|----------------------------------------|--------|
|
|
| `diffusion_models/minimax_h3_fl2va_bf16.safetensors` | **bf16** (fl2va) | 66 GB |
|
|
| `diffusion_models/minimax_h3_ref2va_bf16.safetensors` | **bf16** (ref2va) | 66 GB |
|
|
| `diffusion_models/minimax_h3_fl2va_int8_convrot.safetensors` | **int8** (fl2va) | 34 GB |
|
|
| `diffusion_models/minimax_h3_ref2va_int8_convrot.safetensors` | **int8** (ref2va) | 34 GB |
|
|
| `diffusion_models/minimax_h3_fl2va_pruned_int8_convrot.safetensors` | **int8 pruned** (fl2va) | 21 GB |
|
|
| `diffusion_models/minimax_h3_ref2va_pruned_int8_convrot.safetensors` | **int8 pruned** (ref2va) | 21 GB |
|
|
| `text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors` | text encoder (Qwen3-VL-32B, nvfp4 AWQ) | 16 GB |
|
|
| `vae/minimax_h3_video_vae_fp16.safetensors` | video VAE fp16 | 5.2 GB |
|
|
| `vae/minimax_h3_audio_vae_fp32.safetensors` | audio VAE fp32 | 0.6 GB |
|
|
|
|
**In progress: none.** All nine files fully downloaded and byte-verified.
|
|
|
|
## Not downloaded (exist upstream in the same repo)
|
|
|
|
| File | Size | Why skipped |
|
|
|----------------------------------------------------------------------------|------------|--------------------------------------------------------------------------------------------|
|
|
| `diffusion_models/minimax_h3_{fl2va,ref2va}_pruned_bf16.safetensors` | 40 GB each | a 4th family (pruned, unquantized) — not in the 3 requested families |
|
|
| `diffusion_models/minimax_h3_{fl2va,ref2va}_pruned_fp8_scaled.safetensors` | 21 GB each | a 5th family — same size class as pruned int8 |
|
|
| `text_encoders/qwen3vl_32b_minimax_h3_bf16.safetensors` | 52 GB | every official template uses the nvfp4_awq TE (now fits — grab if quality A/B wanted) |
|
|
| `text_encoders/qwen3vl_32b_minimax_h3_int8_convrot.safetensors` | 27 GB | same |
|
|
|
|
## History / gotchas
|
|
|
|
- 2026-08-06: the pre-existing manual download of ref2va_pruned_int8 had a
|
|
broken filename (`...safetensors?download=true` from a wget of the HF web
|
|
URL); byte size matched HF exactly, so it was renamed and moved to the
|
|
volume rather than re-downloaded.
|
|
- 2026-08-06: bf16 + fl2va_int8 downloads to the volume failed with
|
|
`Disk quota exceeded` at ~93 GB used → tmpfs overflow scheme added.
|
|
- After any **pod restart**: tmpfs weights are gone; ComfyUI will still boot
|
|
(extra_model_paths tolerates the missing dir) but the bf16/fl2va-int8 entries
|
|
disappear from loader dropdowns until `download-weights-tmpfs.sh` is re-run.
|
|
- 2026-08-06 (later): volume resized to 1 TB → pod reset wiped the container
|
|
disk (ComfyUI reinstalled via `install-minimax-h3.sh`); the three overflow
|
|
files were re-downloaded to the volume. Final state: all weights persistent.
|