Upload acid_v4_saos_e22_s16896.ckpt, which was trained in 3x runs from batch size 2 to 4 to 8, from the base SAO-small model.
Browse files
acid_v4_saos_e22_s16896_model_config.json
ADDED
|
@@ -0,0 +1,134 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "diffusion_cond",
|
| 3 |
+
"sample_size": 354304,
|
| 4 |
+
"sample_rate": 44100,
|
| 5 |
+
"audio_channels": 2,
|
| 6 |
+
"model": {
|
| 7 |
+
"pretransform": {
|
| 8 |
+
"type": "autoencoder",
|
| 9 |
+
"iterate_batch": false,
|
| 10 |
+
"model_half": true,
|
| 11 |
+
"config": {
|
| 12 |
+
"encoder": {
|
| 13 |
+
"type": "oobleck",
|
| 14 |
+
"requires_grad": false,
|
| 15 |
+
"config": {
|
| 16 |
+
"in_channels": 2,
|
| 17 |
+
"channels": 128,
|
| 18 |
+
"c_mults": [1, 2, 4, 8, 16],
|
| 19 |
+
"strides": [2, 4, 4, 8, 8],
|
| 20 |
+
"latent_dim": 128,
|
| 21 |
+
"use_snake": true
|
| 22 |
+
}
|
| 23 |
+
},
|
| 24 |
+
"decoder": {
|
| 25 |
+
"type": "oobleck",
|
| 26 |
+
"config": {
|
| 27 |
+
"out_channels": 2,
|
| 28 |
+
"channels": 128,
|
| 29 |
+
"c_mults": [1, 2, 4, 8, 16],
|
| 30 |
+
"strides": [2, 4, 4, 8, 8],
|
| 31 |
+
"latent_dim": 64,
|
| 32 |
+
"use_snake": true,
|
| 33 |
+
"final_tanh": false
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"bottleneck": {
|
| 37 |
+
"type": "vae"
|
| 38 |
+
},
|
| 39 |
+
"latent_dim": 64,
|
| 40 |
+
"downsampling_ratio": 2048,
|
| 41 |
+
"io_channels": 2
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"conditioning": {
|
| 45 |
+
"configs": [
|
| 46 |
+
{
|
| 47 |
+
"id": "prompt",
|
| 48 |
+
"type": "t5",
|
| 49 |
+
"config": {
|
| 50 |
+
"t5_model_name": "google/t5gemma-b-b-ul2",
|
| 51 |
+
"max_length": 128
|
| 52 |
+
}
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"id": "seconds_total",
|
| 56 |
+
"type": "number",
|
| 57 |
+
"config": {
|
| 58 |
+
"min_val": 0,
|
| 59 |
+
"max_val": 256
|
| 60 |
+
}
|
| 61 |
+
}
|
| 62 |
+
],
|
| 63 |
+
"cond_dim": 768
|
| 64 |
+
},
|
| 65 |
+
"diffusion": {
|
| 66 |
+
"cross_attention_cond_ids": ["prompt", "seconds_total"],
|
| 67 |
+
"global_cond_ids": ["seconds_total"],
|
| 68 |
+
"diffusion_objective": "rectified_flow",
|
| 69 |
+
"distribution_shift_options": {
|
| 70 |
+
"min_length": 256,
|
| 71 |
+
"max_length": 4096
|
| 72 |
+
},
|
| 73 |
+
"type": "dit",
|
| 74 |
+
"config": {
|
| 75 |
+
"io_channels": 64,
|
| 76 |
+
"embed_dim": 1024,
|
| 77 |
+
"depth": 16,
|
| 78 |
+
"num_heads": 8,
|
| 79 |
+
"cond_token_dim": 768,
|
| 80 |
+
"global_cond_dim": 768,
|
| 81 |
+
"transformer_type": "continuous_transformer",
|
| 82 |
+
"attn_kwargs": {
|
| 83 |
+
"qk_norm": "ln"
|
| 84 |
+
}
|
| 85 |
+
}
|
| 86 |
+
},
|
| 87 |
+
"io_channels": 64
|
| 88 |
+
},
|
| 89 |
+
"training": {
|
| 90 |
+
"use_ema": true,
|
| 91 |
+
"log_loss_info": false,
|
| 92 |
+
"pre_encoded": true,
|
| 93 |
+
"timestep_sampler": "trunc_logit_normal",
|
| 94 |
+
"optimizer_configs": {
|
| 95 |
+
"diffusion": {
|
| 96 |
+
"optimizer": {
|
| 97 |
+
"type": "AdamW8bit",
|
| 98 |
+
"config": {
|
| 99 |
+
"lr": 5e-5,
|
| 100 |
+
"betas": [0.9, 0.999],
|
| 101 |
+
"eps": 1e-8,
|
| 102 |
+
"weight_decay": 0.01,
|
| 103 |
+
"percentile_clipping": 100,
|
| 104 |
+
"block_wise": true
|
| 105 |
+
}
|
| 106 |
+
},
|
| 107 |
+
"scheduler": {
|
| 108 |
+
"type": "InverseLR",
|
| 109 |
+
"config": {
|
| 110 |
+
"inv_gamma": 1000000,
|
| 111 |
+
"power": 0.5,
|
| 112 |
+
"warmup": 0.995
|
| 113 |
+
}
|
| 114 |
+
}
|
| 115 |
+
}
|
| 116 |
+
},
|
| 117 |
+
"demo": {
|
| 118 |
+
"demo_every": 768,
|
| 119 |
+
"demo_steps": 100,
|
| 120 |
+
"num_demos": 8,
|
| 121 |
+
"demo_cond": [
|
| 122 |
+
{"prompt": "acid lead in F# minor, extremely fast tempo of 200 BPM, 3/4 time signature, 4-bar loop, analog distortion", "seconds_total": 8},
|
| 123 |
+
{"prompt": "drum breaks 174 BPM", "seconds_total": 6},
|
| 124 |
+
{"prompt": "A short, beautiful piano riff in C minor", "seconds_total": 6},
|
| 125 |
+
{"prompt": "Tight Snare Drum", "seconds_total": 1},
|
| 126 |
+
{"prompt": "Glitchy bass design, I used Serum for this", "seconds_total": 4},
|
| 127 |
+
{"prompt": "Synth pluck arp with reverb and delay, 128 BPM", "seconds_total": 6},
|
| 128 |
+
{"prompt": "Acid A minor 120 BPM 4/4 4 bar loop", "seconds_total": 8},
|
| 129 |
+
{"prompt": "Electronic, with a synthesized, futuristic tone. It has a steady, rhythmic pattern and a slightly retro, 1980s-inspired sound. The bass is prominent, giving the track a pulsing, driving feel. The music sets a mood that is upbeat, energetic, and slightly playful. It suggests a setting that could be related to technology, gaming, or a lighthearted, futuristic scenario.", "seconds_total": 8}
|
| 130 |
+
],
|
| 131 |
+
"demo_cfg_scales": [0.5, 1, 1.5, 8]
|
| 132 |
+
}
|
| 133 |
+
}
|
| 134 |
+
}
|