From 9682b0a084e4177cec590f46f3540219a577d755 Mon Sep 17 00:00:00 2001 From: Hanashi Date: Thu, 6 Aug 2026 00:37:44 -0400 Subject: [PATCH] =?UTF-8?q?WEIGHTS.md:=20final=20download=20status=20?= =?UTF-8?q?=E2=80=94=20all=209=20files=20present,=20none=20in=20progress?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-Authored-By: Claude Fable 5 --- docs/WEIGHTS.md | 58 ++++++++++++++++++++++++++++++------------------- 1 file changed, 36 insertions(+), 22 deletions(-) diff --git a/docs/WEIGHTS.md b/docs/WEIGHTS.md index 076109e..9877af2 100644 --- a/docs/WEIGHTS.md +++ b/docs/WEIGHTS.md @@ -1,21 +1,29 @@ # MiniMax H3 — weights log (B200 pod `c306014998a3`, 216.243.220.136) -All files from **[Comfy-Org/MiniMax-H3](https://huggingface.co/Comfy-Org/MiniMax-H3)** (ComfyUI repack of -[MiniMaxAI/MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3)). Two model lines: -**fl2va** (text-to-video + first/last-frame i2v) and **ref2va** (reference-to-video). -ComfyUI finds both storage locations via `/ComfyUI/extra_model_paths.yaml` +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/`. +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 +## Storage locations on the pod | Path | Medium | Notes | |---|---|---| -| `/workspace/minimax-h3/models/` | RunPod network volume | **Persistent.** Volume has a ~100 GB quota — it is full; don't add files without removing others. | -| `/dev/shm/minimax-h3-models/` | tmpfs (RAM, 176 GB) | **EPHEMERAL** — wiped on pod restart. Refill with `bash /workspace/minimax-h3/download-weights-tmpfs.sh`. | +| `/workspace/minimax-h3/models/` | RunPod network volume | **Persistent.** The volume has a ~100 GB quota and is effectively full. | +| `/dev/shm/minimax-h3-models/` | tmpfs (RAM, 176 GB) | **EPHEMERAL** — wiped on pod restart. Refill with `bash /workspace/minimax-h3/download-weights-tmpfs.sh` (~10 min). | -## Downloaded files - -### Persistent — `/workspace/minimax-h3/models/` +## Downloaded — persistent, `/workspace/minimax-h3/models/` | File | Family | Size | |---|---|---| @@ -26,7 +34,7 @@ ComfyUI finds both storage locations via `/ComfyUI/extra_model_paths.yaml` | `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 | -### Ephemeral (tmpfs) — `/dev/shm/minimax-h3-models/` +## Downloaded — ephemeral tmpfs, `/dev/shm/minimax-h3-models/` | File | Family | Size | |---|---|---| @@ -34,24 +42,30 @@ ComfyUI finds both storage locations via `/ComfyUI/extra_model_paths.yaml` | `diffusion_models/minimax_h3_ref2va_bf16.safetensors` | **bf16** (ref2va) | 66 GB | | `diffusion_models/minimax_h3_fl2va_bf16.safetensors` | **bf16** (fl2va) | 66 GB | -The int8 family is split across both locations purely because of the volume -quota: ref2va_int8 landed on the volume before it filled, fl2va_int8 overflowed -to tmpfs. +The int8 family is split across the two locations only because of the volume +quota: ref2va_int8 landed on the volume before it filled; fl2va_int8 and both +bf16 files overflowed to tmpfs. -## Not downloaded (exist upstream) +**In progress: none.** All nine files above are fully downloaded (sizes +byte-verified against the HF API). + +## Not downloaded (exist upstream in the same repo) | File | Size | Why skipped | |---|---|---| -| `diffusion_models/minimax_h3_{fl2va,ref2va}_pruned_bf16.safetensors` | 40 GB each | 4th family (pruned but unquantized); not in the 3 families requested | -| `diffusion_models/minimax_h3_{fl2va,ref2va}_pruned_fp8_scaled.safetensors` | 21 GB each | 5th family; same size class as pruned int8 | -| `text_encoders/qwen3vl_32b_minimax_h3_bf16.safetensors` | 52 GB | nvfp4_awq TE is the one every official template uses | +| `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; bf16 TE would also not fit the volume quota | | `text_encoders/qwen3vl_32b_minimax_h3_int8_convrot.safetensors` | 27 GB | same | ## History / gotchas -- 2026-08-06: original 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 + moved to the volume, not re-downloaded. +- 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. -- Downloads log: `/workspace/minimax-h3/download.log`. +- 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.