Update tokenizer_config.json
#1
by
saattrupdan
- opened
This uses the tokenizer config from NeuralBeagle, with updated model_max_length
to Mistral's 32,768 context length.
saattrupdan
changed pull request status to
closed