ordlibrary/deepsol-clawd-code

1---2license: apache-2.03library_name: transformers4tags:5  - gpt26  - solana7  - clawd8  - code9  - text-generation10language:11  - en12pipeline_tag: text-generation13---14 15# deepsol-clawd-code16 17Merged GPT-2 checkpoint from the Solana Clawd AI training stack (`deepsol-clawd-code-merged`).18 19## Model details20 21| | |22|---|---|23| Architecture | `GPT2LMHeadModel` |24| Layers | 12 |25| Hidden size | 768 |26| Heads | 12 |27| Context | 1024 |28| Vocab | 50257 (GPT-2 tokenizer) |29| Weights dtype | float16 (`model.safetensors`) |30| Size | ~237 MB |31 32## Files33 34- `model.safetensors` — merged weights35- `config.json` / `generation_config.json`36- `tokenizer.json` / `tokenizer_config.json`37 38## Quick start39 40```python41from transformers import AutoModelForCausalLM, AutoTokenizer42 43repo = "ordlibrary/deepsol-clawd-code"44tok = AutoTokenizer.from_pretrained(repo)45model = AutoModelForCausalLM.from_pretrained(repo)46 47prompt = "def transfer_sol("48inputs = tok(prompt, return_tensors="pt")49out = model.generate(**inputs, max_new_tokens=64)50print(tok.decode(out[0], skip_special_tokens=True))51```52 53## Intended use54 55Research / experimentation around Solana-oriented code and tooling assistants in the Clawd training pipeline. This is a small GPT-2-scale model, not a production 7B+ coder.56 57## Limitations58 59- Small capacity vs modern LLMs; expect weak long-context and complex reasoning.60- Training data and merge recipe are project-internal; evaluate before any production use.61- Do not rely on it for financial advice or unsigned transaction construction without human review.62 63## Citation64 65```bibtex66@misc{deepsol-clawd-code,67  title = {deepsol-clawd-code},68  author = {ordlibrary},69  year = {2026},70  howpublished = {\url{https://huggingface.co/ordlibrary/deepsol-clawd-code}}71}72```73