Model to study the differences in batch size

Created: May 28, 2024
Model to study the differences in batch size

This model was created with batch sizes of 4 and 50, each trained for 200 epochs using the same dataset. The purpose of this model is to find evidence or changes related to generalization failure caused by the batch size. Those conducting tests are encouraged to freely alternate between the two models and, if you discover clear differences between the models, please post images or audio files in this thread. This would be greatly beneficial to many.

Model Links -
Batch 4 - https://huggingface.co/SeoulStreamingStation/TestForAIHUB/resolve/main/Test_Batchsize4.zip?download=true

Batch 50 -
https://huggingface.co/SeoulStreamingStation/TestForAIHUB/resolve/main/Test_Batchsize50.zip?download=true

Additional Details:

Tags: No tags available

Download Link: https://huggingface.co/SeoulStreamingStation/TestForAIHUB/resolve/main/Test_Batchsize4.zip%3Fdownload%3Dtrue?download=true Hosted on Hugging Face

Convert this model in EasyAIVoice

Original Link: https://huggingface.co/SeoulStreamingStation/TestForAIHUB/resolve/main/Test_Batchsize4.zip?download=true (404 not found)

Auto-resolved Link: https://huggingface.co/SeoulStreamingStation/TestForAIHUB/resolve/main/Test_Batchsize4.zip%3Fdownload%3Dtrue?download=true

Sample deck

Pitch previews

Listen across generated pitch samples, then Run. Hold Ctrl or Cmd to select multiple pitches.

samples ready
female base 路 pitch 0 Generated female reference sample
Run
male
female
female-base generated sample male-base generated sample Enrichment pending: no scores or voice accords yet Ctrl/Cmd-click = select multiple pitches