WEIGHTS.md: final download status — all 9 files present, none in progress

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

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
+Status as of **2026-08-06 04:35 UTC**.
-[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).
+## Source
-ComfyUI finds both storage locations via `/ComfyUI/extra_model_paths.yaml`
+
+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
+## 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. |
+| `/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`. |
+| `/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
+## Downloaded — persistent, `/workspace/minimax-h3/models/`
-
-### 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
+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 overflowed
+quota: ref2va_int8 landed on the volume before it filled; fl2va_int8 and both
-to tmpfs.
+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_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 | 5th family; same size class as pruned int8 |
+| `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 | nvfp4_awq TE is the one every official template uses |
+| `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
+- 2026-08-06: the pre-existing manual download of ref2va_pruned_int8 had a
-  filename (`...safetensors?download=true` from a wget of the HF web URL); byte
+  broken filename (`...safetensors?download=true` from a wget of the HF web
-  size matched HF exactly so it was renamed + moved to the volume, not re-downloaded.
+  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.