Instructions to use HelioAI/LFM3-1.7B-inference-modules-experiments with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use HelioAI/LFM3-1.7B-inference-modules-experiments 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 HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

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 HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

Use Docker

docker model run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

Llama new
LM Studio
Jan
Ollama
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Ollama:
```
ollama run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
```
Unsloth Desktop

How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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": "HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Docker Model Runner:
```
docker model run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
```

Lemonade

How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

Run and chat with the model

lemonade run user.LFM3-1.7B-inference-modules-experiments-Q5_K_M

List all available models

lemonade list

Hermes Agent

How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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 HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use HelioAI/LFM3-1.7B-inference-modules-experiments with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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 "HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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"

LFM3-1.7B-inference-modules-experiments / experimental /README.md

HelioAI

LFM3-1.7B inference modules: experimental release

67ebcef verified 9 days ago

preview code

Raw History Blame Contribute Delete

2.93 kB

# Experimental checkpoints (modified versions of LFM3-1.7B)

**Notice under the LFM Open License v1.0, section 4(b).** Every file in this folder is a modified version of LiquidAI/LFM3-1.7B. The weights were changed by fine-tuning or by editing the architecture, as described below. The originals are by Liquid AI, Inc. The changes were made in this project. Same license: see `../LICENSE`.

None of these checkpoints beat the unmodified model with the inference modules significantly. They are published so that the negative results can be reproduced. For actual use, take `../weights/` together with `lfm3_deliberate.py` / `lfm3_server.py`.

## Fine-tuned checkpoints

LoRA was trained and merged into bf16. The full-precision version is in `finetunes_hf/<name>/` (transformers format with `train_log.json`); the llama.cpp version is in `finetunes_gguf/<name>-Q5_K_M.gguf`. The numbers are paired changes against the base model on our dev sets; the full context is in `research/REPORT.md`, experiment journal rows 4–5, 16, 19–20, 24.

| Name | What was trained | Data | Result |
|---|---|---|---|
| `iter1` | First GRPO-like RL iteration (LoRA r32, lr 3e-5) | 64 math problems × 8 samples, answer check | 47 of 64 groups gave no signal; effect 0 |
| `bc1`, `bc2` | "Bootstrap of decisiveness": RL groups including forced stops of truncated reasoning as positive examples | 260 / 257 sequences | bc1 +1/140; bc2 −9/140 (biased to negatives) |
| `teach1` | SFT on the shortest correct solution or a checked early stop per problem (LoRA r64, 2 epochs) | 180 own solutions | without modules +3.6 [+0.6, +6.9]; with the module −0.3; MMLU-Pro −4.3 |
| `rft1s` | SFT on correct solutions of unstable problems | 147 own solutions | with the module: math +1.4, BBH +1.1, MMLU-Pro ≈ 0 |
| `rft1rl` | RL by groups (KL 0.02, LoRA r32) | 135 groups, 537 sequences | math +0.8, BBH +3.7, MMLU-Pro ≈ −3 |
| `rft2s` | SFT on correctness with a hard curriculum | 271 own solutions | all changes within noise (−2.5 … +2.9) |
| `rft2rl` | RL by groups with a hard curriculum | 259 groups, 1033 sequences | all changes within noise |

## Architecture edits (GGUF, Q5_K_M)

| File | Change | Result |
|---|---|---|
| `architecture/LFM3-1.7B-loop-layers-10-13-Q5_K_M.gguf` | layers 10–13 repeated once (recurrent depth without training) | math 62 → 40, rejected |
| `architecture/LFM3-1.7B-swa1024-Q5_K_M.gguf` | sliding-attention window 256 → 1024 | not evaluated |
| `architecture/LFM3-1.7B-swa8192-Q5_K_M.gguf` | sliding-attention window 256 → 8192 | not evaluated |

## Latent-thinking runner input

`latent/lfm3-q5km.embd.f32` is the token-embedding table of the Q5_K_M GGUF, dequantized to float32 (128000 × 2048, raw little-endian). It is the input of `research/cpp/soft_think.cpp`, which runs Soft Thinking, Gumbel, Mixture of Inputs and MTI on llama.cpp. You can regenerate it with `research/scripts/dump_embd.py`.

	# Experimental checkpoints (modified versions of LFM3-1.7B)

	Notice under the LFM Open License v1.0, section 4(b). Every file in this folder is a modified version of LiquidAI/LFM3-1.7B. The weights were changed by fine-tuning or by editing the architecture, as described below. The originals are by Liquid AI, Inc. The changes were made in this project. Same license: see `../LICENSE`.

	None of these checkpoints beat the unmodified model with the inference modules significantly. They are published so that the negative results can be reproduced. For actual use, take `../weights/` together with `lfm3_deliberate.py` / `lfm3_server.py`.

	## Fine-tuned checkpoints

	LoRA was trained and merged into bf16. The full-precision version is in `finetunes_hf/<name>/` (transformers format with `train_log.json`); the llama.cpp version is in `finetunes_gguf/<name>-Q5_K_M.gguf`. The numbers are paired changes against the base model on our dev sets; the full context is in `research/REPORT.md`, experiment journal rows 4–5, 16, 19–20, 24.

	\| Name \| What was trained \| Data \| Result \|
	\|---\|---\|---\|---\|
	\| `iter1` \| First GRPO-like RL iteration (LoRA r32, lr 3e-5) \| 64 math problems × 8 samples, answer check \| 47 of 64 groups gave no signal; effect 0 \|
	\| `bc1`, `bc2` \| "Bootstrap of decisiveness": RL groups including forced stops of truncated reasoning as positive examples \| 260 / 257 sequences \| bc1 +1/140; bc2 −9/140 (biased to negatives) \|
	\| `teach1` \| SFT on the shortest correct solution or a checked early stop per problem (LoRA r64, 2 epochs) \| 180 own solutions \| without modules +3.6 [+0.6, +6.9]; with the module −0.3; MMLU-Pro −4.3 \|
	\| `rft1s` \| SFT on correct solutions of unstable problems \| 147 own solutions \| with the module: math +1.4, BBH +1.1, MMLU-Pro ≈ 0 \|
	\| `rft1rl` \| RL by groups (KL 0.02, LoRA r32) \| 135 groups, 537 sequences \| math +0.8, BBH +3.7, MMLU-Pro ≈ −3 \|
	\| `rft2s` \| SFT on correctness with a hard curriculum \| 271 own solutions \| all changes within noise (−2.5 … +2.9) \|
	\| `rft2rl` \| RL by groups with a hard curriculum \| 259 groups, 1033 sequences \| all changes within noise \|

	## Architecture edits (GGUF, Q5_K_M)

	\| File \| Change \| Result \|
	\|---\|---\|---\|
	\| `architecture/LFM3-1.7B-loop-layers-10-13-Q5_K_M.gguf` \| layers 10–13 repeated once (recurrent depth without training) \| math 62 → 40, rejected \|
	\| `architecture/LFM3-1.7B-swa1024-Q5_K_M.gguf` \| sliding-attention window 256 → 1024 \| not evaluated \|
	\| `architecture/LFM3-1.7B-swa8192-Q5_K_M.gguf` \| sliding-attention window 256 → 8192 \| not evaluated \|

	## Latent-thinking runner input

	`latent/lfm3-q5km.embd.f32` is the token-embedding table of the Q5_K_M GGUF, dequantized to float32 (128000 × 2048, raw little-endian). It is the input of `research/cpp/soft_think.cpp`, which runs Soft Thinking, Gumbel, Mixture of Inputs and MTI on llama.cpp. You can regenerate it with `research/scripts/dump_embd.py`.