add total number of runs to logs & small fixes #72
Conversation
src/parameter_search.ipynb
Outdated
| " 'chunk_size' : [192, 256, 320],\n", | ||
| " 'chunk_overlap' : [64, 128, 192],\n", | ||
| " 'chunk_size' : [128, 192, 256, 320],\n", | ||
| " 'chunk_overlap' : [32, 64, 96, 128, 160, 192],\n", |
There was a problem hiding this comment.
is there a risk that this will run for a very long time?
I think the best balance might be in providing few values with large intervals between them e.g. [32, 128, 192] in order to show the contrast in the final score with a relatively short time to complete. We want to minimize the time of wait + API costs for the end user
src/parameter_search.ipynb
Outdated
| "}\n", | ||
| "searchable_shared_params = {\n", | ||
| " 'embed_model' : [(\"huggingface\",\"all-MiniLM-L6-v2\")]\n", | ||
| " 'embed_model' : [(\"huggingface\",\"all-MiniLM-L6-v2\"),(\"huggingface\", \"BAAI/bge-base-en-v1.5\")]\n", |
There was a problem hiding this comment.
why are we re-introducing the bge model again?
There was a problem hiding this comment.
will add it as a comment
src/parameter_search.ipynb
Outdated
| @@ -417,6 +419,7 @@ | |||
| " search_method = search_method,\n", | |||
| " target_metric = target_metric,\n", | |||
| " evaluation_set_path=evaluation_set_path,\n", | |||
| " debug=True,\n", | |||
There was a problem hiding this comment.
should be false I suppose, or do we leave it on for the end user
src/utils.py
Outdated
| check_weaviate_class_exists( | ||
| weaviate_client, | ||
| indexing_config["weaviate_class"], | ||
| ) | ||
|
|
||
| logger.info( |
There was a problem hiding this comment.
This should probable be above line 190 before we start the run
| ) | ||
|
|
||
| logger.info( |
There was a problem hiding this comment.
Same comment as above
No description provided.