udpate ties_merging doc

config/example_config.yaml

@@ -7,4 +7,5 @@ defaults:
 
 fast_dev_run: false # Run a single batch of data to test the model or method
 print_config: true # Print the configuration to the console
+save_report: false # path to save the result report
 

docs/algorithms/ties_merging.md

@@ -1,5 +1,10 @@
 # Ties Merging
 
+<figure markdown="span">
+  ![Image title](images/ties_merging.jpg){ width="750" }
+  <figcaption>Ties-Merging. Credit to <sup id="fnref:1"><a class="footnote-ref" href="#fn:1">1</a></sup></figcaption>
+</figure>
+
 Ties-Merging[^1] represents a novel and structured approach to consolidating multiple task-specific models into a single, efficient multi-task model. This method employs a sequence of deliberate steps to systematically merge task vectors, ensuring that the final model effectively integrates the strengths of each individual task-specific model and resolves potential conflicts between them.
 
 The Ties-Merging algorithm operates through three primary steps:

docs/index.md

@@ -7,7 +7,7 @@
 !!! note
 
     - Any questions or comments can be directed to the [GitHub Issues](https://github.com/tanganke/fusion_bench/issues) page for this project.
-    - Any contributions or pull requests are welcome.
+    - Any contributions or pull requests are welcome. If you find any mistakes or have suggestions for improvements, please feel free to raise an issue or submit a pull request.
 
 
 ## Introduction to Deep Model Fusion

fusion_bench/method/simple_average.py

@@ -77,4 +77,4 @@ def fuse(self, modelpool: ModelPool):
 
         sd = state_dict_mul(sd, 1 / len(modelpool.model_names))
         forward_model.load_state_dict(sd)
-        return simple_average(forward_model)
+        return forward_model

fusion_bench/scripts/cli.py

@@ -11,6 +11,7 @@
 from omegaconf import DictConfig, OmegaConf
 from rich import print as rich_print
 from rich.syntax import Syntax
+import json
 
 from ..method import load_algorithm_from_config
 from ..modelpool import load_modelpool_from_config
@@ -32,7 +33,12 @@ def run_model_fusion(cfg: DictConfig):
 
     if hasattr(cfg, "taskpool") and cfg.taskpool is not None:
         taskpool = load_taskpool_from_config(cfg.taskpool)
-        taskpool.evaluate(merged_model)
+        report = taskpool.evaluate(merged_model)
+        if cfg.get("save_report", False):
+            # save report (Dict) to a file
+            # if the directory of `save_report` does not exists, create it
+            os.makedirs(os.path.dirname(cfg.save_report), exist_ok=True)
+            json.dump(report, open(cfg.save_report, "w"))
     else:
         print("No task pool specified. Skipping evaluation.")