Create README.md

README.md

@@ -0,0 +1,108 @@
+---
+license: other
+license_name: lfm1.0
+license_link: https://huggingface.co/LiquidAI/LFM2.5-230M/blob/main/LICENSE
+language:
+  - en
+base_model: LiquidAI/LFM2.5-230M
+datasets:
+  - DedeProGames/Andy-4.1-NanoAndy
+library_name: transformers
+pipeline_tag: text-generation
+tags:
+  - minecraft
+  - mindcraft
+  - mindcraft-ce
+  - agent
+  - lfm2
+  - edge
+  - conversational
+---
+
+![NanoAndy-230M](logo.png)
+
+# NanoAndy-230M
+
+**NanoAndy-230M** is a compact, full-fine-tuned Minecraft agent model built on [LiquidAI/LFM2.5-230M](https://huggingface.co/LiquidAI/LFM2.5-230M) — Liquid AI's smallest LFM2.5 checkpoint. It is inspired by [Andy-4.1](https://huggingface.co/datasets/Mindcraft-CE/Andy-4.1) and purpose-built to run as the "brain" of a bot in [Mindcraft-CE](https://github.com/mindcraft-ce/mindcraft-ce), the community fork of the open-source Mindcraft platform that lets LLMs control Minecraft characters via Mineflayer.
+
+At only 230M parameters, this is the smallest model in the NanoAndy line — aimed at the tightest memory and compute budgets: low-end GPUs, CPU-only machines, single-board computers, or running many bots side by side.
+
+## What makes it "Nano"
+
+NanoAndy-230M is trained on a stripped-down version of Andy-4.1's conversational data ([DedeProGames/Andy-4.1-NanoAndy](https://huggingface.co/datasets/DedeProGames/Andy-4.1-NanoAndy)):
+
+- **No chain-of-thought.** All `<think>...</think>` reasoning traces were removed from the assistant turns. At 230M parameters there's essentially no spare capacity for long internal monologue, so training goes straight to the final in-game response.
+- **No function-calling turns.** Conversations that used external `tool`-role calls were dropped entirely, keeping the model focused on Mindcraft's native chat/command format instead of a JSON tool-calling schema it wouldn't reliably use at this size.
+
+The result is a lean, fast, direct-response model rather than a smaller reasoning model.
+
+## Model Details
+
+| | |
+|---|---|
+| **Base model** | [LiquidAI/LFM2.5-230M](https://huggingface.co/LiquidAI/LFM2.5-230M) |
+| **Architecture** | LFM2 (hybrid conv + attention) |
+| **Parameters** | ~230M (229,693,184) |
+| **Fine-tuning method** | Full fine-tune (no LoRA/adapters) |
+| **Context length** | 11,264 tokens |
+| **Language** | English |
+| **License** | [LFM Open License v1.0](https://huggingface.co/LiquidAI/LFM2.5-230M/blob/main/LICENSE) (inherited from base model) |
+
+## Training
+
+| | |
+|---|---|
+| **Dataset** | [DedeProGames/Andy-4.1-NanoAndy](https://huggingface.co/datasets/DedeProGames/Andy-4.1-NanoAndy) (1,695 conversations) |
+| **Framework** | [Unsloth](https://github.com/unslothai/unsloth) |
+| **Hardware** | 1x NVIDIA T4 (Google Colab) |
+| **Epochs** | 2 |
+| **Effective batch size** | 8 (1 x 8 grad. accumulation) |
+| **Learning rate** | 5e-5, cosine schedule, 15 warmup steps |
+| **Optimizer** | adamw_8bit |
+| **Final train loss** | ~0.21 |
+
+## Usage
+
+NanoAndy-230M is meant to be dropped into a Mindcraft-CE [bot profile](https://github.com/mindcraft-ce/mindcraft-ce) (e.g. `andy.json`) as the chat/coding model, served locally through something like LM Studio, llama.cpp, or vLLM. Its small footprint makes it a good fit for running on modest hardware or alongside several other bots at once.
+
+It also works with standard `transformers`:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
+
+model_id = "DedeProGames/NanoAndy-230M"
+model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", dtype="bfloat16")
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+
+prompt = "You are a minecraft bot named Andy. A player asks you to gather 4 oak logs."
+input_ids = tokenizer.apply_chat_template(
+    [{"role": "user", "content": prompt}],
+    add_generation_prompt=True,
+    return_tensors="pt",
+    tokenize=True,
+)["input_ids"].to(model.device)
+
+model.generate(
+    input_ids,
+    do_sample=True,
+    temperature=0.3,
+    repetition_penalty=1.05,
+    max_new_tokens=256,
+    streamer=streamer,
+)
+```
+
+## Limitations
+
+- **Not a reasoning model.** With chain-of-thought training data removed, NanoAndy-230M won't show its work — it goes straight to an action/response.
+- **No native tool-calling.** It was not trained on function-call syntax; it expects Mindcraft's native command/chat format.
+- **Very small model.** At 230M parameters — the smallest in the NanoAndy line — it will struggle with long-horizon planning, complex builds, and multi-step reasoning more than the 350M variant, and considerably more than full-size Andy-4 models.
+- **English only.**
+- Narrowly tuned for the Mindcraft agent format — not intended as a general-purpose assistant.
+
+## Acknowledgements
+
+- [Liquid AI](https://liquid.ai) for the LFM2.5 base model.
+- [Mindcraft-CE](https://github.com/mindcraft-ce/mindcraft-ce) and the Andy-4.1 dataset authors for the source conversational data and the platform this model targets.
+- [Unsloth](https://github.com/unslothai/unsloth) for the training tooling.