القائمة
Upload model weights to HuggingFace at high speed using best practices. Use when pushing models, uploading safetensors, uploading GGUFs, or creating model repos on HuggingFace.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | uploading-hf-models |
| description | Upload model weights to HuggingFace at high speed using best practices. Use when pushing models, uploading safetensors, uploading GGUFs, or creating model repos on HuggingFace. |
# Install fast transfer backends
uv pip install hf_transfer hf_xet
# Must be logged in
huggingface-cli login
# or set HF_TOKEN env var
export HF_HUB_ENABLE_HF_TRANSFER=1 # Rust multi-connection uploads
export HF_HUB_DISABLE_XET=1 # Disable xet backend (has stalling bug as of April 2026)
Without hf_transfer, uploads use a single-threaded Python path at ~3 MB/s.
With it: limited by your upstream bandwidth.
Note (April 2026): The hf_xet backend has a known bug causing uploads >2GB to stall progressively. Disable it with HF_HUB_DISABLE_XET=1 until fixed. See HF forum thread.
For large models (>10GB), upload_file per-shard is more reliable than upload_folder or upload_large_folder. Each file gets its own commit — if the upload dies you don't lose progress.
No single LFS file can exceed 50GB. Split safetensors shards to stay under this limit.
Push private, verify, then flip to public. Avoids shipping broken weights.
Every repo needs proper YAML frontmatter for discoverability:
---
base_model: google/original-model-id
pipeline_tag: text-generation
library_name: transformers
language:
- en
license: apache-2.0
tags:
- your-tags
---
For GGUF quantization repos, add:
base_model: your-username/your-bf16-repo
base_model_relation: quantized
Group related models (bf16 + GGUF + variants) into a HF Collection for visibility.
Transformers save_pretrained() creates .cache/ dirs — filter these out before uploading.
If using LoRA/PEFT, call model.merge_and_unload() before save_pretrained(). Otherwise tensor names contain base_layer.weight and lora_ prefixes that break GGUF conversion.
Always save as .safetensors, never .bin or .pth. Safer and faster.
A standalone upload script is at ~/.local/bin/hf-upload:
hf-upload models/my-model username/repo-name # private by default
hf-upload models/my-model username/repo-name --public # public
hf-upload models/my-model username/repo-name --dry-run # list files only
HF_HUB_ENABLE_HF_TRANSFER=1 nohup hf-upload models/my-model user/repo > upload.log 2>&1 &
echo "PID: $! — tail -f upload.log"
HF_HUB_ENABLE_HF_TRANSFER=1. The Python path uses a single HTTP connection that times out on large files.ValueError: cannot update files under .cache/: Filter out .cache/ dirs created by transformers.base_layer.weight tensor names: Call model.merge_and_unload() before saving if using LoRA/PEFT.upload_folder. Each file commits independently.max_shard_size="20GB" in save_pretrained() to split.Sync this Claude Code config against new release notes — find config features added since the pinned baseline that this config should adopt. Use when Claude Code has been updated, when the user asks what's new / what changed / whether to adopt new settings or hooks, after a version bump, or when checking config currency against the latest release. Reports adoptable settings/hooks/env-vars/permissions as a risk-tiered proposal, then bumps the baseline. Do NOT use for auditing existing config quality, debugging hooks, or organizing rules — that is `maintaining-claude-code`.
Monitor and watch GitHub Actions CI/CD pipeline runs in real time. Use when the user has pushed code and wants to watch the build, check CI status, see if tests passed, monitor a pipeline, or wait for a workflow run to complete. Also use after any `jj git push` or `git push` command. Reports pass/fail with failed job logs. Do NOT use for writing or editing CI workflow YAML files, optimizing CI config, or debugging CI configuration — only for monitoring active runs.
Reorganize and clean up messy commit history on a feature branch into logical, reviewable commits. Use when cleaning up commits, preparing a branch for review, separating formatting from logic changes, fixing broken intermediate states, or squashing WIP commits.
Audit and improve Claude Code hooks, rules, and settings.json. Use when adding/debugging a hook, organizing rules, auditing settings.json (permissions, env vars, stale flags), or deciding between hook vs skill vs rule vs CLAUDE.md. For SKILL.md authoring use the skill-creator plugin; for CLAUDE.md audits use the claude-md-improver plugin.
Execute a structured test plan and emit a concise results document (JSON + Markdown). Use when running test plans, executing QA workflows, verifying acceptance criteria systematically, or producing test results documentation.
Search past Claude Code sessions to recall prior solutions, commands, fixes, and decisions. Use when the user references earlier work ("how did I fix", "last time we", "pick up where we left off", "a few months ago", "what was that … I set up", "didn't I once").