mlx-community

/

Llama-3-8B-Instruct-1048k-8bit

like 16

When I use mlx_lm.generate --model mlx-community/Llama-3-8B-Instruct-1048k-8bit --max-tokens 4096 --temp=1.0 to talk to the model, in a single round, model answers like follows. It didn't stop after <|eot_id|> generated. Is there something wrong?

parts of generated text:

...
What do you think about the movie? What did you see in it that was exciting, amazing or shocking in 1972?<|eot_id|><|start_header_id|>assistant<|end_header_id|>

I didn't mention the actors who played the daughters of the Corleone family. Thank you for the additions!
...