HaiwenXia's picture
Upload folder using huggingface_hub
5eb0aae verified
DEVICES: '6'
accelerate:
mixed_precision: 'no'
basics:
random_seed: 42
save_dir: /data/yrb/musicarena/Haiwen/offline_data/cmi-arena/experiments/reward_model
dataset:
cache_dir: /data/yrb/musicarena/Haiwen/offline_data/cmi-arena/tmp
comparison_index: null
db_path: /data/yrb/musicarena/Haiwen/offline_data/cmi-arena/backend/database.db
duration: 30
max_samples: 1000
max_val_samples: null
preference_file: /data/yrb/musicarena/Haiwen/offline_data/cmi-arena/CMI-Training/cleaned_qwen_2025-12-11_15-23-32.json
sample_rate: 24000
val_comparison_index: null
loss:
filter_ties: true
label_smoothing: 0.0
reduction: mean
model:
attention_mode: CA
attn_dropout: 0.0
dim: 768
dim_head: 64
downsample:
factor: 4
kind: mean
ff_dropout: 0.0
ff_mult: 4
freeze_audio: true
freeze_text: true
heads: 8
joint_tf_depth: 4
mlp_dim: 768
mode: concat_text_late
model_name: OpenMuQ/MuQ-MuLan-large
output_dim: 2
prompt_tf_depth: 1
sr: 24000
use_layer_idx: -1
project_root: /data/yrb/musicarena/Haiwen/offline_data/cmi-arena
run_name: debug_downsample
train:
batch_size: 8
betas:
- 0.9
- 0.99
ema_decay: 0.9999
ema_update_every: 1
enable_gradient_checkpointing: true
force_clear_prev_results: false
grad_accum_every: 2
log_tensorboard: true
lr: 1.0e-05
lr_schedule:
min_lr_ratio: 1.0e-06
name: linear_cosine
total_steps: null
warmup_steps: 300
max_grad_norm: 0.5
num_train_steps: 20000
num_valid_batches: 10
num_workers: 8
resume:
checkpoint_path: null
load_optimizer: true
use: false
save_model_every: 1000
use_ema: true
use_lion: false
valid_batch_size: 8
valid_every: 100
valid_frac: 0.1
validate:
checkpoint: null
dir: null
max_failure_cases: 30
num_batches: null
only: false
trust_checkpoint: true