kevinhyj commited on
Commit
a2962bc
·
1 Parent(s): 9b4b5b0

docs: align model card with PhaseFlow_training_data style

Browse files

Add banner figure, related-resource cross-links, and file inventory metadata.

.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ figures/*.png filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -1,34 +1,75 @@
1
  ---
 
2
  license: apache-2.0
3
  library_name: pytorch
4
  tags:
5
  - protein
6
  - peptide
7
- - llps
8
- - dpr
 
9
  ---
10
 
11
- # PhaseFlow checkpoint
12
 
13
- This repository provides the PhaseFlow runtime checkpoint used for peptide
 
 
14
  phase prediction, full-length protein LLPS prediction, and droplet-promoting
15
  region (DPR) localization.
16
 
17
- ## Files
18
-
19
- | File | Description | SHA256 |
20
  | --- | --- | --- |
21
- | `PhaseFlow.pt` | Combined PhaseFlow runtime checkpoint (peptide, full-protein, and DPR weights) | `7fb0091e6dd5a85bd3a6be7a0b606501700c4b8f28ff9b6e309267835a2fdff0` |
 
 
22
 
23
- Download it with:
24
 
25
  ```bash
26
  hf download yanjiehuang/PhaseFlow PhaseFlow.pt --local-dir artifacts/models
27
  ```
28
 
29
- The checkpoint is loaded by the public PhaseFlow codebase. It contains the
30
- embedded peptide PhaseFlow state, the full-protein CondenGT state, and the DPR
31
- stack state; it is not a peptide-only checkpoint.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
32
 
33
- See the [PhaseFlow GitHub repository](https://github.com/kevinhyj/PhaseFlow)
34
- for installation and inference workflows.
 
 
1
  ---
2
+ pretty_name: PhaseFlow
3
  license: apache-2.0
4
  library_name: pytorch
5
  tags:
6
  - protein
7
  - peptide
8
+ - liquid-liquid-phase-separation
9
+ - disordered-protein-regions
10
+ - pytorch
11
  ---
12
 
13
+ # PhaseFlow
14
 
15
+ ![PhaseFlow: sequence data to phase-behaviour learning](figures/PhaseFlow.png)
16
+
17
+ This repository hosts the unified PhaseFlow runtime checkpoint used for peptide
18
  phase prediction, full-length protein LLPS prediction, and droplet-promoting
19
  region (DPR) localization.
20
 
21
+ | Component | Task | Embedded weights |
 
 
22
  | --- | --- | --- |
23
+ | Peptide PhaseFlow | Short-peptide 4×4 PSSI phase-diagram prediction and phase-conditioned sequence generation | Peptide tokenizer, shared Transformer, and Flow Matching stack |
24
+ | CondenGT / LLPS | Protein-level liquid-liquid phase-separation prediction | Full-protein sequence-context encoder and LLPS readout |
25
+ | DPR stack | Driving phase-separation region localization | Region-scoring head and post-processing state for DPR scanning |
26
 
27
+ Download the checkpoint with:
28
 
29
  ```bash
30
  hf download yanjiehuang/PhaseFlow PhaseFlow.pt --local-dir artifacts/models
31
  ```
32
 
33
+ ## Related resources
34
+
35
+ | Resource | Link | Role |
36
+ | --- | --- | --- |
37
+ | PhaseFlow code | [github.com/kevinhyj/PhaseFlow](https://github.com/kevinhyj/PhaseFlow) | Installation, inference scripts, configs, and documentation |
38
+ | PhaseFlow training data | [yanjiehuang/PhaseFlow_training_data](https://huggingface.co/datasets/yanjiehuang/PhaseFlow_training_data) | Public peptide, LLPS, and DPR research-data packages |
39
+ | PhaseFlow web app | [phaseflow.bio](http://phaseflow.bio/) | Online model usage (coming soon; site currently under review) |
40
+
41
+ ## Open access
42
+
43
+ The checkpoint and accompanying model card in this repository are fully open
44
+ source and freely available to the community. We warmly welcome their use in
45
+ research, benchmarking, method development, and reproducible studies. Please cite
46
+ the relevant PhaseFlow work and retain appropriate source attribution when
47
+ using or redistributing the weights.
48
+
49
+ ## Integrity
50
+
51
+ This repository contains `metadata/file_inventory.csv`. Verify all listed files
52
+ before use; the inventory intentionally excludes itself to avoid a
53
+ self-referential checksum.
54
+
55
+ ```python
56
+ import csv
57
+ import hashlib
58
+ from pathlib import Path
59
+
60
+ repo = Path(".")
61
+ with (repo / "metadata/file_inventory.csv").open(newline="") as handle:
62
+ for item in csv.DictReader(handle):
63
+ digest = hashlib.sha256((repo / item["path"]).read_bytes()).hexdigest()
64
+ assert digest == item["sha256"], item["path"]
65
+ ```
66
+
67
+ | File | Description | SHA256 |
68
+ | --- | --- | --- |
69
+ | `PhaseFlow.pt` | Combined PhaseFlow runtime checkpoint (peptide, full-protein, and DPR weights) | `7fb0091e6dd5a85bd3a6be7a0b606501700c4b8f28ff9b6e309267835a2fdff0` |
70
+
71
+ ## Responsible use
72
 
73
+ These weights are research resources, not clinical or diagnostic tools. Model
74
+ predictions require independent validation and must be interpreted under the
75
+ documented evidence and coordinate conventions in the PhaseFlow codebase.
figures/PhaseFlow.png ADDED

Git LFS Details

  • SHA256: 3a278d4878daf84e1544d400a444cffe34143cea10abcf05f4daa5d37c7e7557
  • Pointer size: 132 Bytes
  • Size of remote file: 2.33 MB
metadata/file_inventory.csv ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ path,size_bytes,sha256
2
+ PhaseFlow.pt,340953234,7fb0091e6dd5a85bd3a6be7a0b606501700c4b8f28ff9b6e309267835a2fdff0