dogtooth commited on
Commit
cf2c76f
·
verified ·
1 Parent(s): ba2d889

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +2 -1
README.md CHANGED
@@ -30,7 +30,7 @@ trained with a knowledge cutoff of **January 2017**, from the
30
  ## Usage
31
 
32
  ```python
33
- from transformers import AutoModelForCausalLM, AutoTokenizer
34
  import torch
35
 
36
  model = AutoModelForCausalLM.from_pretrained(
@@ -43,6 +43,7 @@ model = AutoModelForCausalLM.from_pretrained(
43
  ## Conversion Notes
44
 
45
  - Converted from the original Open LM `.pt` checkpoint to HuggingFace `LlamaForCausalLM` format.
 
46
  - The original model uses **QK norm** (RMSNorm on Q and K projections), which is not natively
47
  supported by HF Llama. QK norm weights are dropped during conversion. For exact numerical
48
  equivalence, use the [open_lm](https://github.com/mlfoundations/open_lm) library.
 
30
  ## Usage
31
 
32
  ```python
33
+ from transformers import AutoModelForCausalLM
34
  import torch
35
 
36
  model = AutoModelForCausalLM.from_pretrained(
 
43
  ## Conversion Notes
44
 
45
  - Converted from the original Open LM `.pt` checkpoint to HuggingFace `LlamaForCausalLM` format.
46
+ - Architecture dimensions are auto-detected from checkpoint weights.
47
  - The original model uses **QK norm** (RMSNorm on Q and K projections), which is not natively
48
  supported by HF Llama. QK norm weights are dropped during conversion. For exact numerical
49
  equivalence, use the [open_lm](https://github.com/mlfoundations/open_lm) library.