mirror of https://github.com/storytold/cloud-worker.git synced 2026-10-09 00:09:43 +00:00

Files

T

Hanashi 225aa97620 MiniMax H3 on RunPod: install/download scripts, ComfyUI wiring, benchmark harness, research docs

- scripts/pod: weight downloads (workspace + tmpfs overflow for volume quota),
  extra_model_paths.yaml, headless ComfyUI launcher, test image generator,
  one-shot pod installer
- scripts/local: auto-reconnecting port forward for the ComfyUI panel
- bench: API-based harness for t2v/i2v/ref2v across durations, resolutions,
  and weight families; CSV/JSONL results
- docs: WEIGHTS.md (what's downloaded where), RESEARCH.md (ComfyUI guides,
  no-Comfy options via SGLang/vLLM/diffusers, concurrency model, serverless)

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

2026-08-06 00:28:16 -04:00

3.0 KiB

Raw Blame History

MiniMax H3 — weights log (B200 pod `c306014998a3`, 216.243.220.136)

All files from Comfy-Org/MiniMax-H3 (ComfyUI repack of 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 (installed from scripts/pod/extra_model_paths.yaml).

Storage locations

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`.

Downloaded files

Persistent — `/workspace/minimax-h3/models/`

File	Family	Size
`diffusion_models/minimax_h3_ref2va_pruned_int8_convrot.safetensors`	int8 pruned (ref2va)	21 GB
`diffusion_models/minimax_h3_fl2va_pruned_int8_convrot.safetensors`	int8 pruned (fl2va)	21 GB
`diffusion_models/minimax_h3_ref2va_int8_convrot.safetensors`	int8 (ref2va)	34 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

Ephemeral (tmpfs) — `/dev/shm/minimax-h3-models/`

File	Family	Size
`diffusion_models/minimax_h3_fl2va_int8_convrot.safetensors`	int8 (fl2va)	34 GB
`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.

Not downloaded (exist upstream)

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
`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: 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.

3.0 KiB Raw Blame History

MiniMax H3 — weights log (B200 pod c306014998a3, 216.243.220.136)