Instructions to use ramankrishna10/npc-reason with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ramankrishna10/npc-reason with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ramankrishna10/npc-reason",
	filename="npc-reason-f16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ramankrishna10/npc-reason 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 ramankrishna10/npc-reason:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ramankrishna10/npc-reason:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ramankrishna10/npc-reason:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ramankrishna10/npc-reason: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 ramankrishna10/npc-reason:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ramankrishna10/npc-reason: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 ramankrishna10/npc-reason:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ramankrishna10/npc-reason:Q4_K_M

Use Docker

docker model run hf.co/ramankrishna10/npc-reason:Q4_K_M

LM Studio
Jan

vLLM

How to use ramankrishna10/npc-reason with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ramankrishna10/npc-reason"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ramankrishna10/npc-reason",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ramankrishna10/npc-reason:Q4_K_M

Ollama
How to use ramankrishna10/npc-reason with Ollama:
```
ollama run hf.co/ramankrishna10/npc-reason:Q4_K_M
```

Unsloth Studio

How to use ramankrishna10/npc-reason with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ramankrishna10/npc-reason to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ramankrishna10/npc-reason to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ramankrishna10/npc-reason to start chatting

Atomic Chat new
Docker Model Runner
How to use ramankrishna10/npc-reason with Docker Model Runner:
```
docker model run hf.co/ramankrishna10/npc-reason:Q4_K_M
```

Lemonade

How to use ramankrishna10/npc-reason with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ramankrishna10/npc-reason:Q4_K_M

Run and chat with the model

lemonade run user.npc-reason-Q4_K_M

List all available models

lemonade list

npc-reason / reports /dispatch2_complete.md

ramankrishna10

verifier + eval harness + reports + locks + usage

1ba301a verified about 1 month ago

preview code

Raw

History Blame Contribute Delete

2.41 kB

=== NPC Reason — Dispatch 2: V4 Corpus + SFT Distill ===

Corpus: 13,245 V4 (deepseek-v4-flash) chains from TRAIN splits (GSM8K-train + MATH-train 7 subj) -> mechanically filtered against FROZEN verifier (verifiable AND correct) -> 7,546 kept (keep rate 56.97%). TRAINSET.lock frozen. reject breakdown: no_assertions 3894 (29%, mostly non-arithmetic MATH) · bad_arithmetic 1211 (9%) · non_composing 387 · wrong_answer 206 · dup 1 decontam vs EVAL.lock: 0 overlap (train/test disjoint by construction) Finding: the strong teacher emits ungrounded/wrong chains ~43% of the time on this distribution; all rejected before training (the strict-basis move).

SFT: R1-Distill-1.5B + LoRA r64 (7 modules, 73.9M = 3.99% trainable), batch1 x accum16, 3 epochs, seq 1024 constant-pad (0.03% truncated), sdpa + gradient checkpointing. train loss 0.538 -> 0.075, eval 0.311, runtime 2h31m. Merged model saved (3.5GB).

Result: (format-prompt, vs frozen baseline, greedy, frozen verifier) verifiable 76.8% (base 0.0) · accuracy 65.8% (base 61.6) · verified+correct 58.0% (base 0.0) mean assertions/chain 3.44 (base 0.0)

PREREG: lift +58.0pp verified-and-correct -> SUPPORTED (bar +15pp) verifiable-rate 76.8% -> NOT SUPPORTED (bar 90%) accuracy delta +4.2pp (improved) -> SUPPORTED (bar: no regress >5pp) => 2 of 3 SUPPORTED; all-three gate not met.

RL gate: WARRANTED via the high-signal condition (76.8% verifiable / 58% v&c gives the verifier-reward RL dense signal); all-three bar NOT cleared (verifiable < 90 is the explicit RL target). Not a null — a large real lift missing one of three bars.

Honest: Thesis validated — distilling ONLY verifier-confirmed chains took the model from 0% mechanically-checkable reasoning to 76.8% verifiable / 58% verified-and-correct, with accuracy rising not falling. The grounding transferred; the remaining ~23% unverifiable tail is the RL phase's job.

Decisions np-007..np-010 · all frozen locks intact · eval decontaminated · SFT_RESULT.lock frozen Next: Dispatch 3 = RLVR/GRPO with the frozen verifier as reward, target verifiable-rate >=90%, RL criteria pre-registered against a held-out slice never used for SFT selection.