Fix load model test

Convert model first
Code typos
Add help message to readme

.github/workflows/test.yaml

@@ -49,4 +49,6 @@ jobs:
       # Load target model checkpoint and run inference on example variants
     - name: Load and test target METL model
       shell: bash --login {0}
-      run: python code/tests.py --checkpoint_path output/training_logs/DgLkMZxu/checkpoints/epoch=49-step=50.ckpt --variants E3K,G102S;T36P,S203T,K207R;V10A,D19G,F25S,E113V --dataset avgfp
+      run: |
+        python code/convert_ckpt.py output/training_logs/DgLkMZxu/checkpoints/epoch=49-step=50.ckpt
+        python code/tests.py --checkpoint_path output/training_logs/DgLkMZxu/checkpoints/DgLkMZxu.pt --variants E3K,G102S;T36P,S203T,K207R;V10A,D19G,F25S,E113V --dataset avgfp

.gitignore

@@ -1,3 +1,7 @@
+# Output files
+output/*
+!output/README.md
+
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]

README.md

@@ -53,6 +53,11 @@ python code/train_source_model.py @args/pretrain_avgfp_local.txt --max_epochs 5
 The test metrics are expected to show poor performance after such a short training run.
 For instance, `pearson_total_score` may be around 0.24.
 
+See the help message for an explanation of all the arguments
+```bash
+python code/train_source_model.py --help
+```
+
 # Finetuning on experimental data
 
 Experimental data is stored in [dms_data](data/dms_data) directory. 
@@ -71,3 +76,8 @@ python code/train_target_model.py @args/finetune_avgfp_local.txt --enable_progre
 ```
 
 Following the short demonstration pretraining and finetuning process is expected to give test set Spearman correlation around 0.6.
+
+See the help message for an explanation of all the arguments
+```bash
+python code/code/train_target_model.py --help
+```

code/tests.py

@@ -1,5 +1,4 @@
 """ testing code """
-import argparse
 import metl
 import torch
 import utils
@@ -46,4 +45,4 @@ def load_checkpoint_run_inference(checkpoint_path, variants, dataset):
 
     parsed_args = parser.parse_args()
 
-    load_checkpoint_run_inference(args.checkpoint_path, args.variants, args.dataset)
+    load_checkpoint_run_inference(parsed_args.checkpoint_path, parsed_args.variants, parsed_args.dataset)