h2o-llmstudio
H2O LLM Studio is a framework and graphical user interface developed for the fine-tuning (fine-tuning) processes of large language models. It allows to customize models using techniques such as LoRA and 8-bit training without the need to write code. It also offers the opportunity to visualize model performance, evaluate and export through the Hugging Face Hub.
h2oai · Local AI / Self-Hosted LLM
Common commercial products in this category: ChatGPT, Claude, Gemini
- You can fine-tune large language models through the graphical interface without writing code.
- With LoRA and 8-bit training techniques, you can train models with low memory usage.
- You can easily transfer the models you have trained to the Hugging Face Hub platform.
- ↓ NegativeGitHub Issue
[BUG] Error during LoRA-merge in HF upload for Llama 3.1 70B model
“### 🐛 Bug Today when attempting to upload a LoRA-trained Llama 3.1 70B model (first time I've trained Llama 3.1), I hit the following during the eLoRA merge. Note I used the `cpu_shard` method to upload. I've tried it twice now with the same error. 2024-07-24 17:22:58,705 -…”