TheBloke commited on
Commit
72de88a
1 Parent(s): 4d6b21d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -26,9 +26,9 @@ tags:
26
 
27
  These files are pytorch format fp16 model files for [Elinas' Chronos 33B](https://huggingface.co/elinas/chronos-33b) merged with [Kaio Ken's SuperHOT 30B 8K LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) to produce a model capable of 8K context.
28
 
29
- [Kaio Ken's SuperHOT 30B LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using the provided monkey patch (`llama_rope_scaled_monkey_patch.py`)
30
 
31
- Alternatively, `config.json` can be modified to allow the monkey patch to load via trust_remote_code=True. I plan to update this repo shortly to include that method.
32
 
33
  ## Repositories available
34
 
 
26
 
27
  These files are pytorch format fp16 model files for [Elinas' Chronos 33B](https://huggingface.co/elinas/chronos-33b) merged with [Kaio Ken's SuperHOT 30B 8K LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) to produce a model capable of 8K context.
28
 
29
+ [Kaio Ken's SuperHOT 30B LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
30
 
31
+ Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
32
 
33
  ## Repositories available
34