libc++abi: terminating due to uncaught exception of type std::runtime_error: [Matmul::eval_cpu] Currently only supports float32.

[bengrine]

python run.py --train --model ./TinyLlama-1.1B-Chat-v1.0 --data ./data --batch-size 1 --lora-layers 4

Loading pretrained model You are using the default legacy behaviour of the <class 'transformers.models.llama.tokenization_llama_fast.LlamaTokenizerFast'>. This is expected, and simply means that the legacy(previous) behavior will be used so nothing changes for you. If you want to use the new behaviour, setlegacy=False. This should only be set if you understand what it means, and thoroughly read the reason why this was added as explained in https://github.com/huggingface/transformers/pull/24565 Total parameters 227.461M Trainable parameters 0.205M Loading datasets Training libc++abi: terminating due to uncaught exception of type std::runtime_error: [Matmul::eval_cpu] Currently only supports float32. zsh: abort python run.py --train --model ./TinyLlama-1.1B-Chat-v1.0 --data ./data 1 4

Model Name: Mac Pro
Model Identifier: MacPro7,1
Enclosure: Tower
Processor Name: 16-Core Intel Xeon W
Processor Speed: 3.2 GHz
Number of Processors: 1
Total Number of Cores: 16
L2 Cache (per Core): 1 MB
L3 Cache: 22 MB
Hyper-Threading Technology: Enabled
Memory: 128 GB

[bengrine]

I guess only Apple Silicon is supported in MLX