ShuxianZou commited on
Commit
27e43fe
·
verified ·
1 Parent(s): cb069d2

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +34 -9
README.md CHANGED
@@ -2,8 +2,8 @@
2
  datasets:
3
  - genbio-ai/transcript_isoform_expression_prediction
4
  base_model:
5
- - EleutherAI/enformer-official-rough
6
  - genbio-ai/AIDO.RNA-1.6B-CDS
 
7
  - facebook/esm2_t30_150M_UR50D
8
  metrics:
9
  - spearmanr
@@ -13,6 +13,10 @@ tags:
13
  ---
14
  # Tri-modal model for RNA isoform expression prediction
15
 
 
 
 
 
16
 
17
  ## Model architecture
18
 
@@ -20,13 +24,34 @@ tags:
20
  <img src="https://cdn-uploads.huggingface.co/production/uploads/63008d4bc1e149ceaff724a3/wKWIDdLu418MpXyihbCj9.png" alt="description" style="width:100%; height:auto;">
21
  </p>
22
 
23
- Task:
24
- * Input: dna_seq, rna_seq, protein_seq
25
- * Output: expression level in 30 tissues
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
 
27
- Backbones:
28
- * DNA: Enformer (fully finetuning)
29
- * RNA: AIDO.RNA-1.6B-CDS (lora finetuning)
30
- * Protein: ESM2-150M (lora finetuning)
31
 
32
- Multimodal Fusion: concat fusion
 
 
 
 
 
 
 
2
  datasets:
3
  - genbio-ai/transcript_isoform_expression_prediction
4
  base_model:
 
5
  - genbio-ai/AIDO.RNA-1.6B-CDS
6
+ - EleutherAI/enformer-official-rough
7
  - facebook/esm2_t30_150M_UR50D
8
  metrics:
9
  - spearmanr
 
13
  ---
14
  # Tri-modal model for RNA isoform expression prediction
15
 
16
+ ## RNA isoform expression prediction
17
+ * Input: dna_seq, rna_seq, protein_seq
18
+ * Output: expression level in 30 tissues
19
+
20
 
21
  ## Model architecture
22
 
 
24
  <img src="https://cdn-uploads.huggingface.co/production/uploads/63008d4bc1e149ceaff724a3/wKWIDdLu418MpXyihbCj9.png" alt="description" style="width:100%; height:auto;">
25
  </p>
26
 
27
+ * Backbones:
28
+ * DNA: Enformer (fully finetuning)
29
+ * RNA: AIDO.RNA-1.6B-CDS (lora finetuning)
30
+ * Protein: ESM2-150M (lora finetuning)
31
+ * Fusion method: concat fusion
32
+
33
+
34
+ ## Usage
35
+ **Download model**
36
+ ```python
37
+ from huggingface_hub import snapshot_download
38
+ from pathlib import Path
39
+
40
+ model_name = "genbio-ai/AIDO.MM-Enformer-RNA-1.6B-CDS-ESM2-150M-ConcatFusion-rna-isoform-expression-ckpt"
41
+ genbio_models_path = Path.home().joinpath('genbio_models', model_name)
42
+ genbio_models_path.mkdir(parents=True, exist_ok=True)
43
+ snapshot_download(repo_id=model_name, local_dir=genbio_models_path)
44
+ ```
45
 
46
+ **Evaluation script**
47
+ ```bash
48
+ CONFIG_FILE=... # put the config file path here
49
+ CKPT_PATH=... # put the model checkpoint path here
50
 
51
+ mgen test --config $CONFIG_FILE \
52
+ --data.batch_size 16 \
53
+ --trainer.logger null \
54
+ --model.strict_loading False \
55
+ --model.reset_optimizer_states True \
56
+ --ckpt_path $CKPT_PATH
57
+ ```