Safetensors
model-ingest commited on
Commit
6bd1f49
·
verified ·
1 Parent(s): f87bed3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -18
README.md CHANGED
@@ -13,9 +13,6 @@ license: mit
13
  - **Languages:** SMARTS & SMILES (chemical substructure representation)
14
  - **Pipeline Tag:** text2text-generation for MMP transformation
15
  - **Library:** Transformers, PyTorch
16
-
17
- ---
18
-
19
  # 2. Intended Use
20
 
21
  - **Direct Use:**
@@ -43,18 +40,12 @@ license: mit
43
  - **MMP-C2V:**
44
  - Integration into analog enumeration pipelines
45
  - Integration into high-throughput virtual screening pipelines
46
-
47
- ---
48
-
49
  # 3. Bias, Risks, and Limitations
50
 
51
  - **Known Limitations:** The models rely on the availability and coverage of large historical transformation datasets, and its performance may vary in underrepresented chemical domains.
52
  - **Biases:** Inherits biases from ChEMBL-derived medicinal chemistry literature.
53
  - **Risk Areas:** Our framework is intended for research use and does not introduce specific ethical concerns.
54
  - **Recommendations:** None
55
-
56
- ---
57
-
58
  # 4. Training Details
59
 
60
  - **Training Data:** Raw data is downloaded from ChEMBL database and available at [https://chembl.gitbook.io/chembl-interface-documentation/downloads](https://chembl.gitbook.io/chembl-interface-documentation/downloads).
@@ -72,9 +63,6 @@ license: mit
72
  - Hardware:
73
  - MMPT-FM: 4 × NVIDIA A6000 GPUs
74
  - MMP variants: 4 × NVIDIA H100 GPUs
75
-
76
- ---
77
-
78
  # 5. Evaluation
79
 
80
  - **Metrics:**
@@ -86,16 +74,10 @@ license: mit
86
  - Within-patent analog generation (PMV17)
87
  - Cross-patent analog generation (PMV17 → PMV21)
88
  - **Testing Data:** Patent-derived datasets from PMV Pharmaceuticals (2017, 2021)
89
-
90
- ---
91
-
92
  # 6. Usage
93
 
94
  - **Sample Inference Code:** Described conceptually in the publications; code corresponds to variable-to-variable generation with beam search can be found at our GitHub repository: [https://github.com/MSDLLCpapers/MMPTTransformer](https://github.com/MSDLLCpapers/MMPTTransformer).
95
  - **GitHub Links:** [https://github.com/MSDLLCpapers/MMPTTransformer](https://github.com/MSDLLCpapers/MMPTTransformer)
96
-
97
- ---
98
-
99
  # 7. Citation
100
 
101
  **BibTeX:**
 
13
  - **Languages:** SMARTS & SMILES (chemical substructure representation)
14
  - **Pipeline Tag:** text2text-generation for MMP transformation
15
  - **Library:** Transformers, PyTorch
 
 
 
16
  # 2. Intended Use
17
 
18
  - **Direct Use:**
 
40
  - **MMP-C2V:**
41
  - Integration into analog enumeration pipelines
42
  - Integration into high-throughput virtual screening pipelines
 
 
 
43
  # 3. Bias, Risks, and Limitations
44
 
45
  - **Known Limitations:** The models rely on the availability and coverage of large historical transformation datasets, and its performance may vary in underrepresented chemical domains.
46
  - **Biases:** Inherits biases from ChEMBL-derived medicinal chemistry literature.
47
  - **Risk Areas:** Our framework is intended for research use and does not introduce specific ethical concerns.
48
  - **Recommendations:** None
 
 
 
49
  # 4. Training Details
50
 
51
  - **Training Data:** Raw data is downloaded from ChEMBL database and available at [https://chembl.gitbook.io/chembl-interface-documentation/downloads](https://chembl.gitbook.io/chembl-interface-documentation/downloads).
 
63
  - Hardware:
64
  - MMPT-FM: 4 × NVIDIA A6000 GPUs
65
  - MMP variants: 4 × NVIDIA H100 GPUs
 
 
 
66
  # 5. Evaluation
67
 
68
  - **Metrics:**
 
74
  - Within-patent analog generation (PMV17)
75
  - Cross-patent analog generation (PMV17 → PMV21)
76
  - **Testing Data:** Patent-derived datasets from PMV Pharmaceuticals (2017, 2021)
 
 
 
77
  # 6. Usage
78
 
79
  - **Sample Inference Code:** Described conceptually in the publications; code corresponds to variable-to-variable generation with beam search can be found at our GitHub repository: [https://github.com/MSDLLCpapers/MMPTTransformer](https://github.com/MSDLLCpapers/MMPTTransformer).
80
  - **GitHub Links:** [https://github.com/MSDLLCpapers/MMPTTransformer](https://github.com/MSDLLCpapers/MMPTTransformer)
 
 
 
81
  # 7. Citation
82
 
83
  **BibTeX:**