quarterturn commited on
Commit
c36fe28
1 Parent(s): be71aca

update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -17,12 +17,12 @@ For max precision clone Molmo-7B-D-0924:
17
  ```
18
  You'll need a 24GB GPU since the model loads at bf16.
19
 
20
- For slightly less preccision, but much lower memory needed, clone molmo-7B-D-bnb-4bit:
21
  ```
22
  git lfs install
23
  git clone https://huggingface.co/cyan2k/molmo-7B-D-bnb-4bit
24
  ```
25
- A 12GB GPU should be fine.
26
 
27
  1. create a python3 venv or use conda to create an environment, eg:
28
  ``` conda create -n caption python=3.11 ```
 
17
  ```
18
  You'll need a 24GB GPU since the model loads at bf16.
19
 
20
+ For less precision, but much lower memory needed, clone molmo-7B-D-bnb-4bit:
21
  ```
22
  git lfs install
23
  git clone https://huggingface.co/cyan2k/molmo-7B-D-bnb-4bit
24
  ```
25
+ A 12GB GPU should be fine. Note that the 4-bit quant produces not just less accurate, but quite different in it's description. YMMV.
26
 
27
  1. create a python3 venv or use conda to create an environment, eg:
28
  ``` conda create -n caption python=3.11 ```