transformerlab / transformerlab-app

Experiment with Large Language Models

transformerlab/transformerlab-app Issues

Importing External Dataset system shows 0 datasize even is imported and data is 13 Mbytes
Closed 2 months ago10
On Training no progress indication for queued jobs
Updated 2 months ago2
Flash_attn doesn't build on windows
Updated 2 months ago
Local Connection installer: No error reported if there are exceptions during "Install Dependencies" step
Updated 2 months ago1
Better error when you try to download/start a restricted model without HuggingFace key
Closed 2 months ago2
MacOS-X using mlx - Missing parameters: lm_head.biases lm_head.scales.
Updated 2 months ago1
Vicuna models don't start
Updated 2 months ago
Test each model that can work in MLX
Updated 2 months ago2
Better errors when starting models fails
Closed 2 months ago1
Generate config.json files for exported models
Updated 2 months ago1
If you eject a model while it is running the app can get into a weird state
Closed 2 months ago1
Control where models and other large data is stored
Updated 2 months ago2
Display something next to models that require HuggingFace authentication
Updated 3 months ago
install script requires curl but doesn't check for it
Updated 3 months ago1
Navigating away from the Model Store UI will cancel a model download
Updated 3 months ago1
Need to install Flash Attention after all other requirements
Closed 3 months ago1
Fix code scanning alert - Potentially unsafe external link
Updated 3 months ago
Mac OS using mlx - what is the required "Adaptor Name" field in Plugin Config tab of Training Template?
Closed 3 months ago2
System Prompt display may not always match what is stored
Closed 3 months ago
UI Can go Blank on Crash
Updated 3 months ago
Electron Upgrader has no GUI
Closed 3 months ago2
Unable to install Conda - ERROR: File or directory already exists: '~/.transformerlab/miniconda3'
Closed 3 months ago3
Remember previously used experiment
Closed 3 months ago
Add a regenerate icon to messages
Closed 3 months ago
We do not upgrade dependencies on install
Closed 3 months ago2
Deleted Conda environment doesn't allow fresh install
Closed 3 months ago1
Error running llama trainer if no NVIDIA GPU
Updated 3 months ago
If you reinstall Transformer Lab, Plugins need to be reinstalled
Updated 3 months ago
Llama trainer looking in wrong directory for custom datasets
Closed 3 months ago1
Downloads don't resume or show progress using "Download HuggingFace Model" button
Updated 3 months ago
Add a jobs directory to workspace to store job artifacts
Updated 3 months ago
Training on TinyDolphin model throws errors in MLX training script
Updated 3 months ago1
Preview more of a dataset
Updated 3 months ago
Inference on Trained Mistral-7b fails often
Updated 3 months ago1
Add a "stop" button for inference ("interact" tab)
Updated 3 months ago
Token counter for inference seems buggy
Updated 3 months ago
Add epochs to mlx_lora_trainer
Updated 3 months ago
Export models names should include quantization if there is any
Updated 3 months ago
Make it optional to fuse the adapter/lora
Updated 3 months ago
Error reinstalling mlx_lora_trainer - mlx_examples dependency
Updated 3 months ago
Give detailed instrucitons on writing a template
Updated 3 months ago
More flexible custom dataset file structures
Updated 3 months ago
Have a way to force update of app
Closed 3 months ago1
Pin MLX to some tested version
Closed 3 months ago1
Check mlx_lora_trainer for missing parameters
Closed 3 months ago3
Change standard value for temperature
Closed 3 months ago1
Unable to select any quantization other than 4 for MLX Exporter
Closed 4 months ago1
Models generated by MLX training won't run
Closed 4 months ago1
Select A Model screen doesn't scroll
Closed 4 months ago2
Error when training with very few examples
Closed 4 months ago1