BluScriptCP / logs /in2_train_csv /CP_Training_log9_config.txt
unparallelled's picture
Upload folder using huggingface_hub
d7922f9 verified
Raw History Blame Contribute Delete
561 Bytes
Configuration (bluscriptCP):
d_model: 1536
n_layers: 36
q_heads: 12
kv_heads: 4 (G=3, head_dim=128)
ffn_hidden: 4096
context_length: 4096
vocab_size: 50304 (tying=0)
B: 1
T: 4096
world_size: 1 (dp=1 x cp=1)
CP_ATTN_MODE: ring (load_balance=HeadTail)
global_batch: 8192
grad_accum_steps: 2
Parameters: 1060624896
Parameters per GPU: 1060624896
Max Learning Rate: 0.0006
Min Learning Rate: 6e-05
max_steps: 2
warmup_steps: 1
checkpointing: 1
ckpt_freq: 50
run: 9
GPU Memory Used (rank 0): 9096.9 MB / 143155.9 MB