mathav95raj

pip install talos is failing due to sklearn in requirements.txt instead of scikit-learn

3

#### 1) Confirm the below - [Yes ] My Python version is 3.5 or higher (3.10.10) - [Yes ] I have searched through the [Issues](https://github.com/autonomio/talos/issues) for a duplicate `pip install...

pip install wrangle is failing due to sklearn in requirements.txt instead of scikit-learn

1

1) Confirm the below [Yes ] My Python version is 3.5 or higher (3.10.10) [Yes ] I have searched through the [Issues](https://github.com/autonomio/talos/issues) for a duplicate `pip install wrangle` is failing...

`pip install astetik` is failing because it is unable to install the deprecated sklearn dependency. In the [pypi page](https://pypi.org/project/sklearn/#descriptionl) of sklearn, there is a notice to start using scikit-learn instead...

What is the recommended GPU configuration to chat fine tune the full sequence length of 8192?

2

Even with a NVIDIA A100 80 GB GPU, I am not able to fine tune the model on full sequence length of 8192. I was not able to fine tune...

Bug: Huge difference between prompt processing (tokens/sec) compared to Llama cpp or Ollama

2

### What happened? For llama cpp I had downloaded the q4_k_m quantized [model](https://huggingface.co/jxtngx/Meta-Llama-3.2-1B-Q4_K_M-GGUF/tree/main) and used [llama-bench](https://github.com/ggerganov/llama.cpp/tree/master/examples/llama-bench). For ollama I pulled the q4_k_m model from ollama. By running the model with...

MLX VLM Eval - Paligemma - Failed: keyNotFound(base: "VisionEmbeddings", key: "_positionIds")

1

paligemma model abruptly jumps at the last stage from 90% to 100% and the error appears when loading

Llava models are not consistent in response even after following specified prompt template

1

Model : [llava-llama-3-8b-v1_1-int4.gguf](https://huggingface.co/xtuner/llava-llama-3-8b-v1_1-gguf/blob/main/llava-llama-3-8b-v1_1-int4.gguf) Clip : [llava-llama-3-8b-v1_1-mmproj-f16.gguf](https://huggingface.co/xtuner/llava-llama-3-8b-v1_1-gguf/blob/734381b71afc8644ae87cb50f7872a3c8d43411c/llava-llama-3-8b-v1_1-mmproj-f16.gguf) According to [docs](https://llmfarm.tech/docs/prompt_format , I set up the prompt template as follows, [system](You are a helpful, respectful and honest assistant.) ### User: {prompt} Describe...

KeyError: 'image_token_index' when training with LoRA on Qwen2.5-VL-3B-Instruct-bf16

6

## Description When attempting to fine-tune the Qwen2.5-VL-3B-Instruct-bf16 model using LoRA, I'm encountering a KeyError for 'image_token_index' during training. ## Environment - mlx-vlm version: 0.1.26 - Model: mlx-community/Qwen2.5-VL-3B-Instruct-bf16 - OS:...