Instructions to use frontier-infra/jebadiah-27b-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use frontier-infra/jebadiah-27b-GGUF 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 frontier-infra/jebadiah-27b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Use Docker

docker model run hf.co/frontier-infra/jebadiah-27b-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use frontier-infra/jebadiah-27b-GGUF with Ollama:
```
ollama run hf.co/frontier-infra/jebadiah-27b-GGUF:Q4_K_M
```
Unsloth Desktop

How to use frontier-infra/jebadiah-27b-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

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": "frontier-infra/jebadiah-27b-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use frontier-infra/jebadiah-27b-GGUF with Docker Model Runner:
```
docker model run hf.co/frontier-infra/jebadiah-27b-GGUF:Q4_K_M
```

Lemonade

How to use frontier-infra/jebadiah-27b-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.jebadiah-27b-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use frontier-infra/jebadiah-27b-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

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 frontier-infra/jebadiah-27b-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use frontier-infra/jebadiah-27b-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf frontier-infra/jebadiah-27b-GGUF:Q4_K_M

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 "frontier-infra/jebadiah-27b-GGUF:Q4_K_M" \
  --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"

jbrashear commited on 4 days ago

Commit

e4b70f7

verified ·

1 Parent(s): e653e65

Temperatures: take the 27B's refit (score 0.7558) from frontier-infra/jebadiah-27b@1c0d794f

Browse files

Files changed (2) hide show

README.md +1 -1
temperatures.json +38 -2

README.md CHANGED Viewed

@@ -42,7 +42,7 @@ needs about its own size in GPU or unified memory, plus about 1 GB for a 4k cont
 The answer is the log probability of each option label ("A", "B", ...) at the answer position, which
 llama-server's `/completion` returns. The script renders the prompt exactly as AINode does, sends the raw
 text (so the server's own chat template is never used), renormalises over the labels and applies
-`temperatures.json` (choice 1.2321, noul 1.297, score 1.1423). You need a llama.cpp that knows the `qwen35` architecture: we checked
 v0.5.0 (older builds refuse the file).
 ```bash

 The answer is the log probability of each option label ("A", "B", ...) at the answer position, which
 llama-server's `/completion` returns. The script renders the prompt exactly as AINode does, sends the raw
 text (so the server's own chat template is never used), renormalises over the labels and applies
+`temperatures.json` (choice 1.2321, noul 1.297, score 0.7558). You need a llama.cpp that knows the `qwen35` architecture: we checked
 v0.5.0 (older builds refuse the file).
 ```bash

temperatures.json CHANGED Viewed

@@ -2,9 +2,25 @@
  "temperatures": {
   "choice": 1.2321,
   "noul": 1.297,
-  "score": 1.1423
  },
- "applied_target": "train",
  "calib_file": "/workspace/jeb/data-v1/calib.jsonl",
  "n": {
   "choice": 225,
@@ -45,6 +61,26 @@
     "nll_before": 1.059,
     "nll_after": 1.055
    }
   }
  },
  "nll_before": {

  "temperatures": {
   "choice": 1.2321,
   "noul": 1.297,
+  "score": 0.7558
  },
+ "applied_target": "mixed",
+ "applied_fits": {
+  "choice": "train",
+  "noul": "train",
+  "score": "hard"
+ },
+ "previous": {
+  "applied_target": "train",
+  "temperatures": {
+   "choice": 1.2321,
+   "noul": 1.297,
+   "score": 1.1423
+  },
+  "replaced": "2026-09-26"
+ },
+ "why": "Refit 2026-09-26 without retraining. Both fits are unchanged and come from the calibration split only. On every evaluation set, re-tempering the stored logits: score questions calibrate better at the hard fit (T 0.76) than at the train fit (T 1.14), whose ordinal target is deliberately smoothed; choice and noul stay on the train fit, which the public sets prefer. Question-weighted ECE over the 20 sets 0.0809 -> 0.0718, macro 0.0708 -> 0.0642, NLL 0.5233 -> 0.5139; accuracy unchanged. See eval/RESULTS.md, Temperature fits.",
+ "top_level_stats_describe": "the train fit (nll_before/nll_after/ece_before/ece_after below are its calibration-split numbers)",
  "calib_file": "/workspace/jeb/data-v1/calib.jsonl",
  "n": {
   "choice": 225,
     "nll_before": 1.059,
     "nll_after": 1.055
    }
+  },
+  "mixed": {
+   "choice": {
+    "T": 1.2321,
+    "nll_before": 0.4621,
+    "nll_after": 0.4537,
+    "source": "train"
+   },
+   "noul": {
+    "T": 1.297,
+    "nll_before": 0.3953,
+    "nll_after": 0.387,
+    "source": "train"
+   },
+   "score": {
+    "T": 0.7558,
+    "nll_before": 0.883,
+    "nll_after": 0.8662,
+    "source": "hard"
+   }
   }
  },
  "nll_before": {