Niksa Praljak commited on
Commit
d85f01b
1 Parent(s): 26a45a2

Add section on model weight installation

Browse files
Files changed (1) hide show
  1. README.md +24 -10
README.md CHANGED
@@ -32,11 +32,32 @@ Create and activate a conda environment and install the required packages:
32
  ```bash
33
  conda create -p /env_path/BioM3_env python=3.10 # /env_path/ is the location that contains the conda env
34
  conda activate /env_path/BioM3_env
 
35
  cd /path/BioM3 # /path/ is the location that contains the huggingface repo for BioM3
36
  sh torch_requirements.sh # install torch software
37
  pip install -r requirements.txt # install remaining packages
38
  ```
39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
40
  ## Stage 1: PenCL Inference
41
 
42
  ### Overview
@@ -51,10 +72,9 @@ Before running the model, ensure you have:
51
 
52
  ### Running the Model
53
 
54
- 1. Clone the repository:
55
  ```bash
56
- git clone https://huggingface.co/your_username/BioM3_PenCL
57
- cd BioM3_PenCL
58
  ```
59
 
60
  2. Run inference:
@@ -139,13 +159,7 @@ Before running the model, ensure you have:
139
 
140
  ### Running the Facilitator Model
141
 
142
- 1. Clone the repository:
143
- ```bash
144
- git clone https://huggingface.co/your_username/BioM3_Facilitator
145
- cd BioM3_Facilitator
146
- ```
147
-
148
- 2. Run inference:
149
  ```bash
150
  python run_Facilitator_sample.py \
151
  --json_path "stage2_facilitator_config.json" \
 
32
  ```bash
33
  conda create -p /env_path/BioM3_env python=3.10 # /env_path/ is the location that contains the conda env
34
  conda activate /env_path/BioM3_env
35
+ git clone https://huggingface.co/niksapraljak1/BioM3 /path/
36
  cd /path/BioM3 # /path/ is the location that contains the huggingface repo for BioM3
37
  sh torch_requirements.sh # install torch software
38
  pip install -r requirements.txt # install remaining packages
39
  ```
40
 
41
+ ## Model Weights Installation
42
+
43
+ Before running models, change directory to `BioM3/weights` folder, follow instructions, and download pretrained weights for the desired BioM3 configuration:
44
+
45
+ ```bash
46
+ cd /path/BioM3/weights
47
+ # after changing directory, follow instructions of README.md to install weights for each model component
48
+ ```
49
+
50
+ Note: choose the desired BioM3 configuration/checkpoint, then install weights for each folder:
51
+ - `/path/BioM3/weights/PenCL`
52
+ - `/path/BioM3/weights/Facilitator`
53
+ - `/path/BioM3/weights/ProteoScribe`
54
+
55
+ Each folder contains a `README.md` detailing the different model weight configurations. For benchmarking, the optimal configuration is:
56
+ - `BioM3_PenCL_epoch20.bin`
57
+ - `BioM3_Facilitator_epoch20.bin`
58
+ - `BioM3_ProteoScribe_epoch20.bin`
59
+
60
+
61
  ## Stage 1: PenCL Inference
62
 
63
  ### Overview
 
72
 
73
  ### Running the Model
74
 
75
+ 1. Change directory to BioM3 repo:
76
  ```bash
77
+ cd /path/BioM3 # /path/ where is the location to the cloned BioM3 repo
 
78
  ```
79
 
80
  2. Run inference:
 
159
 
160
  ### Running the Facilitator Model
161
 
162
+ 1. Run sampling:
 
 
 
 
 
 
163
  ```bash
164
  python run_Facilitator_sample.py \
165
  --json_path "stage2_facilitator_config.json" \