h2o-llmstudio
h2o-llmstudio copied to clipboard
h2oai
H2O LLM Studio - a framework and no-code GUI for fine-tuning LLMs. Documentation: https://docs.h2o.ai/h2o-llmstudio/
### 🚀 Feature Add model card and sample code for inference to the Download model feature ### Motivation UX
### 🚀 Feature Select device that should be used to merge weights (for Download and for push to Hugging Face) ### Motivation CPU might be slow for larger models
https://platform.openai.com/docs/guides/rate-limits/error-mitigation
The setting `Mask Prompt Labels` allows to fully mask the prompt labels and only calculate the loss on the output. When chaining conversation data while training, we will have a...
### 🚀 Feature Soweit ich sehen konnte, gibt es derzeit kein Bedienfeld, mit dem ich die angesprochenen Environment Variable editieren kann, was dazu führt, dass das Statement WARNING: `More than...
### 🚀 Feature Add more inference parameters: - Top P - Top K
### 🐛 Bug Hi there! I'm working on getting native support h2o-llmstudio on RunPod containers. I was able to get it fully made into docker container that includes training downloads...
This PR adds default lora target layers for falcon models. I have excluded MPT models, as they seem to require additional code changes.
### 🚀 Feature With https://github.com/h2oai/h2o-llmstudio/issues/78 being merged, we may want to consider training own reward models within LLM Studio. This should probably be a new task type and requires a...