What is the maximum number of generated stories that `rasa visualize` can deal with?

Rasa Open Source
1

Hi everyone,

What is the maximum number of generated stories that rasa visualize can deal with? I have quite a few checkpoints (maybe too many, we might need to generate our data differently). Right now, the rasa visualize hangs indefinitely.

Here is the debug message if that helps:

rasa visualize --debug
2019-08-22 11:18:09 DEBUG    pykwalify.compat  - Using yaml library: /home/nicolas/miniconda3/envs/chatbot/lib/python3.6/site-packages/ruamel/yaml/__init__.py
2019-08-22 11:18:10 DEBUG    rasa.nlu.training_data.loading  - Training data format of '/tmp/tmp1glkfrnu/374ddd8d4af44315929d3141074178bb_nlu.md' is 'md'.
2019-08-22 11:18:10 DEBUG    rasa.nlu.training_data.loading  - Training data format of '/tmp/tmp1glkfrnu/72671a26f6d54451ad98805d918ac7d1_files.md' is 'md'.
2019-08-22 11:18:10 INFO     rasa.core.visualize  - Starting to visualize stories...
2019-08-22 11:18:10 DEBUG    rasa.core.training.generator  - Number of augmentation rounds is 0
2019-08-22 11:18:10 DEBUG    rasa.core.training.generator  - Starting data generation round 0 ... (with 1 trackers)
Processed Story Blocks: 100%|██████████████████████████████████████████| 76/76 [00:53<00:00,  8.19s/it, # trackers=7644]
2019-08-22 11:19:03 DEBUG    rasa.core.training.generator  - Finished phase (8832 training samples found).
2019-08-22 11:19:03 DEBUG    rasa.core.training.generator  - Data generation rounds finished.
2019-08-22 11:19:03 DEBUG    rasa.core.training.generator  - Found 0 unused checkpoints
2019-08-22 11:19:03 DEBUG    rasa.core.training.generator  - Found 8832 training trackers.

Thanks for your help! Nicolas

2

Did it ever finish? There is no hard limit, but the deduplication heuristic to simplify the graph is quite costly. Even if it finishes, if you have that many stories, you won’t see anything in the resulting graph.

What do you want to achieve - maybe there is another way?