mlx-community/Kimi-VL-A3B-Thinking-4bit

Upload tokenizer_config.json with huggingface_hub

Upload tokenizer_config.json with huggingface_hub

Delete tokenization_moonshot.py

Delete tiktoken.model

Upload tokenizer_config.json with huggingface_hub

Upload tokenizer.json with huggingface_hub

Delete tokenizer_config.json with huggingface_hub

Delete tokenizer.json with huggingface_hub

Upload tokenizer_config.json with huggingface_hub

Add <|im_assistant|> to eos_token_id: [163584, 163586]

Set eos_token_id=163584 in config.json

Set eos_token_id=163584 for <|im_end|>

Set eos_token to <|im_end|> for proper chat termination

Fix media_placeholder_token_id to match converted tokenizer

Add pre-converted tokenizer_config.json (no tiktoken dependency)

Add pre-converted tokenizer.json (no tiktoken dependency)

Delete modeling_kimi_vl.py

Upload 15 files

Upload folder using huggingface_hub

Delete tiktoken.model

Upload folder using huggingface_hub

Add files using upload-large-folder tool

initial commit