Instructions to use eyad-silx/progress with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use eyad-silx/progress with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf eyad-silx/progress # Run inference directly in the terminal: llama cli -hf eyad-silx/progress
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf eyad-silx/progress # Run inference directly in the terminal: llama cli -hf eyad-silx/progress
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf eyad-silx/progress # Run inference directly in the terminal: ./build/bin/llama cli -hf eyad-silx/progress
Use Docker
docker model run hf.co/eyad-silx/progress
- Llama new
- LM Studio
- Jan
- Ollama
How to use eyad-silx/progress with Ollama:
ollama run hf.co/eyad-silx/progress
- Unsloth Desktop
- Pi
How to use eyad-silx/progress with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf eyad-silx/progress
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "eyad-silx/progress" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use eyad-silx/progress with Docker Model Runner:
docker model run hf.co/eyad-silx/progress
- Lemonade
How to use eyad-silx/progress with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull eyad-silx/progress
Run and chat with the model
lemonade run user.progress-{{QUANT_TAG}}List all available models
lemonade list
- Hermes Agent
How to use eyad-silx/progress with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf eyad-silx/progress
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default eyad-silx/progress
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use eyad-silx/progress with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf eyad-silx/progress
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "eyad-silx/progress" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Download artifacts/scripts/identity_check.py from eyad-silx/progress: direct link, hf CLI and curl.
- Browser
- Download file 3.55 kB
-
https://huggingface.co/eyad-silx/progress/resolve/main/artifacts/scripts/identity_check.py
- Command line
-
hf download hf://eyad-silx/progress/artifacts/scripts/identity_check.py
-
curl -L -o identity_check.py https://huggingface.co/eyad-silx/progress/resolve/main/artifacts/scripts/identity_check.py
3.55 kB
| """Identity round-trip sanity v2 (chunked): rebuild PQ2_0 scales from the king's own | |
| weights (codes unchanged) along the correct block axis; write back as a challenger.""" | |
| import hashlib | |
| import shutil | |
| import sys | |
| from pathlib import Path | |
| sys.path.insert(0, "/root/ingot/src") | |
| import numpy as np | |
| from common.gguf import read_header | |
| from common.gguf.types import PQ2_0 | |
| KING = Path("/root/models/king-ix/Ingot-IX-27B-Q2.gguf") | |
| OUT_DIR = Path("/root/models/chal-identity") | |
| OUT = OUT_DIR / KING.name | |
| CHUNK = 1 << 20 # blocks per chunk | |
| g = read_header(KING) | |
| pq = [(name, t) for name, t in g.tensors.items() if t.ggml_type == PQ2_0] | |
| print(f"PQ2_0 tensors: {len(pq)}") | |
| def codes_d(blocks): | |
| d = blocks[:, :2].copy().view("<f2").astype(np.float32).reshape(-1) | |
| qs = blocks[:, 2:] | |
| codes = np.stack([(qs >> (2 * k)) & 3 for k in range(4)], axis=-1).reshape(-1, 128).astype(np.float32) - 1.0 | |
| return codes, d | |
| def refit_scales(blocks): | |
| codes, d = codes_d(blocks) | |
| den = (codes * codes).sum(1) | |
| num = (codes * codes * d[:, None]).sum(1) | |
| return np.where(den > 0, num / np.maximum(den, 1e-9), d) | |
| # ---- Part A: axis diagnostic on small PQ2_0 tensors ---- | |
| print("=== AXIS DIAGNOSTIC ===") | |
| shown = 0 | |
| for name, t in pq: | |
| n_blocks = int(np.prod(t.shape)) // 128 | |
| if n_blocks > 200_000: | |
| continue | |
| b = np.asarray(np.memmap(KING, np.uint8, "r", g.data_start + t.offset, (n_blocks * 34,)).reshape(-1, 34)) | |
| d_orig = b[:, :2].copy().view("<f2").astype(np.float32).reshape(-1) | |
| d_ok = refit_scales(b).astype(np.float16).astype(np.float32) | |
| err = np.abs(d_ok - d_orig) | |
| scale = max(np.abs(d_orig).max(), 1e-9) | |
| # v1's wrong axis on a square grid, for the failure record | |
| side = int(round(np.sqrt(n_blocks))) | |
| bad = None | |
| if side * side == n_blocks: | |
| codes, _ = codes_d(b) | |
| dT = d_orig.reshape(side, side).T.reshape(-1) | |
| den = (codes * codes).sum(1) | |
| num = (codes * codes * dT[:, None]).sum(1) | |
| d_bad = np.where(den > 0, num / np.maximum(den, 1e-9), dT).astype(np.float16).astype(np.float32) | |
| bad = (np.abs(d_bad - d_orig).max() / scale) | |
| print(f" {name[:46]:<46} blocks={n_blocks:>8} maxerr={err.max()/scale:.2e} bits_flipped={(d_ok.astype('<f2').view(np.uint16) != d_orig.astype('<f2').view(np.uint16)).sum()}" | |
| + (f" v1-axis maxerr={bad:.2e}" if bad is not None else "")) | |
| shown += 1 | |
| if shown >= 8: | |
| break | |
| # ---- Part B: write identity challenger ---- | |
| OUT_DIR.mkdir(parents=True, exist_ok=True) | |
| shutil.copy(KING, OUT) | |
| patched = changed = 0 | |
| with open(OUT, "r+b") as fh: | |
| for name, t in pq: | |
| n_blocks = int(np.prod(t.shape)) // 128 | |
| raw = np.memmap(OUT, np.uint8, "r+", g.data_start + t.offset, (n_blocks * 34,)) | |
| for s in range(0, n_blocks, CHUNK): | |
| e = min(s + CHUNK, n_blocks) | |
| b = np.asarray(raw[s * 34:e * 34]).reshape(-1, 34) | |
| d_new = refit_scales(b).astype(np.float16).astype("<f2").view(np.uint8).reshape(-1, 2) | |
| changed += int((d_new != b[:, :2]).sum() // 2) | |
| b[:, :2] = d_new | |
| patched += 1 | |
| if patched % 100 == 0: | |
| print(f" patched {patched}/{len(pq)}", flush=True) | |
| print(f"\npatched {patched} tensors; scales differing from original: {changed}") | |
| def h(p): | |
| m = hashlib.sha256() | |
| with open(p, "rb") as f: | |
| for c in iter(lambda: f.read(1 << 24), b""): | |
| m.update(c) | |
| return m.hexdigest()[:16] | |
| print(f"king sha256 {h(KING)}") | |
| print(f"identity sha256 {h(OUT)}") | |
| print("WROTE", OUT) | |