Instructions to use Maelstrome/lora-wave-session-r32 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Maelstrome/lora-wave-session-r32 with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("unsloth/gemma-4-e2b-it-unsloth-bnb-4bit")
model = PeftModel.from_pretrained(base_model, "Maelstrome/lora-wave-session-r32")

llama-cpp-python

How to use Maelstrome/lora-wave-session-r32 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Maelstrome/lora-wave-session-r32",
	filename="gguf/gemma-4-e2b-it-peft.Q4_K_M-00001-of-00005.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Maelstrome/lora-wave-session-r32 with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Maelstrome/lora-wave-session-r32:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Maelstrome/lora-wave-session-r32:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Maelstrome/lora-wave-session-r32:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Maelstrome/lora-wave-session-r32: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 Maelstrome/lora-wave-session-r32:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Maelstrome/lora-wave-session-r32: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 Maelstrome/lora-wave-session-r32:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Maelstrome/lora-wave-session-r32:Q4_K_M

Use Docker

docker model run hf.co/Maelstrome/lora-wave-session-r32:Q4_K_M

LM Studio
Jan

vLLM

How to use Maelstrome/lora-wave-session-r32 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Maelstrome/lora-wave-session-r32"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Maelstrome/lora-wave-session-r32",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Maelstrome/lora-wave-session-r32:Q4_K_M

Ollama
How to use Maelstrome/lora-wave-session-r32 with Ollama:
```
ollama run hf.co/Maelstrome/lora-wave-session-r32:Q4_K_M
```

Unsloth Studio new

How to use Maelstrome/lora-wave-session-r32 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 Maelstrome/lora-wave-session-r32 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 Maelstrome/lora-wave-session-r32 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Maelstrome/lora-wave-session-r32 to start chatting

Docker Model Runner
How to use Maelstrome/lora-wave-session-r32 with Docker Model Runner:
```
docker model run hf.co/Maelstrome/lora-wave-session-r32:Q4_K_M
```

Lemonade

How to use Maelstrome/lora-wave-session-r32 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Maelstrome/lora-wave-session-r32:Q4_K_M

Run and chat with the model

lemonade run user.lora-wave-session-r32-Q4_K_M

List all available models

lemonade list

Maelstrome commited on 15 days ago

Commit

31c1f72

verified ·

1 Parent(s): 3a226e2

Upload report/README.md with huggingface_hub

Browse files

Files changed (1) hide show

report/README.md +12 -8

report/README.md CHANGED Viewed

@@ -14,9 +14,11 @@ datasets:
 - Maelstrome/lora-wave-session-dataset
 ---
-# lora-wave-session-r32 — training report
-Documentation-only repo. Contains the full training/eval write-up for the **rank-32 / 1-epoch A100** WAVE fine-tune of Gemma 4 E2B Instruct, plus the head-to-head comparison against its rank-16 sibling.
 ## Documents
@@ -26,13 +28,15 @@ Documentation-only repo. Contains the full training/eval write-up for the **rank
 | [`COMPARISON.md`](./COMPARISON.md) | Head-to-head vs the rank-16 / 3-epoch sibling run (`lora-wave-session`). Same dataset, same seed, same test split. r32 wins on every probability metric. |
 | [`MORNING_REPORT.md`](./MORNING_REPORT.md) | First-pass overnight summary written immediately after training completed. Preserved for history; superseded by `REPORT.md`. |
-## Linked artifacts
-- 🦥 **Adapter:** [`Maelstrome/lora-wave-session-r32`](https://huggingface.co/Maelstrome/lora-wave-session-r32) — PEFT adapter (~194 MB)
-- 🛻 **Merged bf16:** [`Maelstrome/lora-wave-session-r32-merged`](https://huggingface.co/Maelstrome/lora-wave-session-r32-merged) — drop-in for `transformers`/vLLM (~10 GB)
-- 🧊 **GGUF Q4_K_M:** [`Maelstrome/lora-wave-session-r32-gguf`](https://huggingface.co/Maelstrome/lora-wave-session-r32-gguf) — llama.cpp / Ollama / wllama (~4 GB)
-- 📚 **Dataset:** [`Maelstrome/lora-wave-session-dataset`](https://huggingface.co/datasets/Maelstrome/lora-wave-session-dataset) — 4,277 examples, frozen splits (seed `7`)
-- 🌊 **Sibling run (rank-16):** [`Maelstrome/lora-wave-session`](https://huggingface.co/Maelstrome/lora-wave-session) — same dataset, different recipe
 ## Headline numbers

 - Maelstrome/lora-wave-session-dataset
 ---
+# `report/` — Training & evaluation write-up
+The full training/eval documentation for the **rank-32 / 1-epoch A100** WAVE fine-tune of Gemma 4 E2B Instruct, plus the head-to-head comparison against its rank-16 sibling.
+Originally a standalone repo (`Maelstrome/lora-wave-session-r32-report`); now lives as a subdirectory of the consolidated [`Maelstrome/lora-wave-session-r32`](https://huggingface.co/Maelstrome/lora-wave-session-r32) repo alongside the adapter weights, GGUF, and MediaPipe artifacts.
 ## Documents
 | [`COMPARISON.md`](./COMPARISON.md) | Head-to-head vs the rank-16 / 3-epoch sibling run (`lora-wave-session`). Same dataset, same seed, same test split. r32 wins on every probability metric. |
 | [`MORNING_REPORT.md`](./MORNING_REPORT.md) | First-pass overnight summary written immediately after training completed. Preserved for history; superseded by `REPORT.md`. |
+## Sibling artifacts in this repo
+- 🦥 **PEFT adapter** — at the [repo root](../) (~194 MB). Pairs with `unsloth/gemma-4-E2B-it`.
+- 🧊 **GGUF Q4_K_M (5-shard split)** — [`gguf/`](../gguf) (~3.2 GB). For llama.cpp / Ollama / [wllama](https://github.com/ngxson/wllama) browser runtime.
+- 📱 **MediaPipe LiteRT bundle** — [`mediapipe/`](../mediapipe) (~4.95 GB). For MediaPipe LLM Inference (Android / iOS / web).
+- 📚 **Dataset:** [`Maelstrome/lora-wave-session-dataset`](https://huggingface.co/datasets/Maelstrome/lora-wave-session-dataset) — 4,277 examples, frozen splits (seed `7`).
+- 🌊 **Sibling rank-16 run:** [`Maelstrome/lora-wave-session`](https://huggingface.co/Maelstrome/lora-wave-session) — same dataset, different recipe (rank-16 / 3-epoch RTX 5080). Has the same subdir layout.
+> The `-merged` and `-gguf` sibling repos referenced in older versions of this doc were consolidated and deleted. The current `gguf/` subdir is a fresh build from a PEFT re-merge (the original unsloth-merged base produced corrupt all-`<pad>` output and is no longer published).
 ## Headline numbers