You can deploy this model and start earning money today!

soda-research/
discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0

Name: discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0
Availability: InStock

next-token-prediction

isoflop

research

apache-2.0

Views

No views yet

Discrete Audio IsoFLOP Model (discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0)

A suite of discrete audio models trained for our IsoFLOP study as part of SODA, which is a unified next-token prediction on interleaved semantic, acoustic, and text tokens.

🥤 Project Page: https://soda-audio.github.io

For full usage instructions (e.g., inference code), and more information, please refer to the SODA-4B-base model card.

The details for this particular model is as follows:

compute_budget: 3e20
param_count (non-embedding): 4.24B
hidden_dim: 2816
num_layers: 27
batch_size: 64
training_step: 42769
hash_key: b829f0

📈 WandB: https://wandb.ai/potsawee/marin/groups/IsoFlop/workspace

soda-research/discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0

Discrete Audio IsoFLOP Model (discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0)

soda-research/
discrete-audio-isoflop-3e20-4.24B-d2816-L27-B64-b829f0