Kernel crashed while loading checkpoint shards
Hi there, just a question I don't understand why my kernel crashed when I run :
tokenizer = LlamaTokenizer.from_pretrained("mistral-7B-v0.1_extract")
model = MistralForCausalLM.from_pretrained("mistral-7B-v0.1_extract")
Note : mistral_7B-v0.1_extract is th efolder where I extracted the content from the raw checkpoints.
If I use "mistralai/Mistral-7B-v0.1" directly I have the same issue.
This is the output : {
"stack": "The Kernel crashed while executing code in the the current cell or a previous cell. Please review the code in the cell(s) to identify a possible cause of the failure."
}
I would make sure you have safetensros
and that you have enough ram to load the model / check for OOM errors
your RAM is not enough even i faced this issue.
Is there a way to get around this? (apart from buying a better machine)