saumya-pailwan commited on
Commit
c51d47d
·
verified ·
1 Parent(s): 8c8b390

description and hints update

Browse files
Files changed (1) hide show
  1. app.py +14 -6
app.py CHANGED
@@ -155,8 +155,8 @@ model_card = ModelCard(
155
  name="TRIA: The Rhythm In Anything",
156
  description=(
157
  "Transform your rhythmic ideas into full drum performances. TRIA takes two short audio prompts: \n "
158
- "Rhythm Prompt (tapping, beatboxing, or percussion gesture) "
159
- "and a Timbre Prompt (an example drum sound or kit recording) \n "
160
  "It generates 3 drum arrangements that match your groove and chosen timbre. "
161
  ),
162
  author="Patrick O'Reilly, Julia Barnett, Hugo Flores García, Annie Chu, Nathan Pruyne, Prem Seetharaman, Bryan Pardo",
@@ -173,10 +173,18 @@ with gr.Blocks(title="TRIA") as demo:
173
  model_dropdown = gr.Dropdown(choices=model_names, value=model_names[0], label="Model")
174
 
175
  with gr.Row():
176
- cfg_scale = gr.Slider(0.0, 10.0, value=2.0, step=0.1, label="CFG Scale")
177
- top_p = gr.Slider(0.0, 1.0, value=0.95, step=0.01, label="Top P")
178
- mask_temperature = gr.Slider(0.0, 20.0, value=10.5, step=0.1, label="Mask Temperature")
179
- seed = gr.Slider(0, 1000, value=0, step=1, label="Random Seed")
 
 
 
 
 
 
 
 
180
 
181
  out1 = gr.Audio(type="filepath", label="Generated #1")
182
  out2 = gr.Audio(type="filepath", label="Generated #2")
 
155
  name="TRIA: The Rhythm In Anything",
156
  description=(
157
  "Transform your rhythmic ideas into full drum performances. TRIA takes two short audio prompts: \n "
158
+ "Timbre Prompt: an example recording for the desired sound (e.g. drum sound) \n "
159
+ "Rhythm Prompt: the sound gesture expressing the desired pattern (e.g. tapping or beatboxing) \n"
160
  "It generates 3 drum arrangements that match your groove and chosen timbre. "
161
  ),
162
  author="Patrick O'Reilly, Julia Barnett, Hugo Flores García, Annie Chu, Nathan Pruyne, Prem Seetharaman, Bryan Pardo",
 
173
  model_dropdown = gr.Dropdown(choices=model_names, value=model_names[0], label="Model")
174
 
175
  with gr.Row():
176
+ cfg_scale = gr.Slider(0.0, 10.0, value=2.0, step=0.1, label="CFG Scale", info=(
177
+ "Controls how strongly the model follows your prompts/conditions.\n"
178
+ "Low values: Model is more creative/random, High values: Model adheres strictly to prompts, less variation."))
179
+ top_p = gr.Slider(0.0, 1.0, value=0.95, step=0.01, label="Top P", info=(
180
+ "Nucleus Sampling: Limits the pool of possible next tokens to the most likely ones.\n"
181
+ "(Default behav.) only sample from tokens that make up the top 95% of probability"))
182
+ mask_temperature = gr.Slider(0.0, 20.0, value=10.5, step=0.1, label="Mask Temperature", info=(
183
+ "In masked models, controls the randomness of which tokens get unmasked first.\n"
184
+ "Low temperature: More deterministic, predictable unmasking order, High temperature: More random unmasking, can add variety"))
185
+ seed = gr.Slider(0, 1000, value=0, step=1, label="Random Seed", info=(
186
+ "Same seed and same inputs: Produces identical outputs (reproducible).\n"
187
+ "Set a specific seed to get consistent results, or use fifferent seeds to try variations"))
188
 
189
  out1 = gr.Audio(type="filepath", label="Generated #1")
190
  out2 = gr.Audio(type="filepath", label="Generated #2")