Skip to main content
zhehuaic
Projects
canary-v0_speechllm
Reports
use_am_tokenizer
Log in
Sign up
Share
Comment
Star
Share
Comment
Star
use_am_tokenizer
Zhehuai Chen
Created on March 19
|
Last edited on March 19
Comment
/lustre/fs8/portfolios/llmservice/users/zhehuaic/works/mod_speech_llm/scripts/oci_canary_gpt_llama_tiny_crossbmg4eghel_sp_lhmain3cross.sh
/lustre/fs8/portfolios/llmservice/users/zhehuaic/works/mod_speech_llm/scripts/oci_canary_gpt_llama_tiny_crossbmg4eghel_sa_lhmain3cross.sh
Section 1
train_loss
train_loss
Select runs that logged train_loss
to visualize data in this line chart.
train_step_timing in s
train_step_timing in s
5k
10k
15k
20k
25k
global_step
3
3.5
4
4.5
5
5.5
6
reduced_train_loss
reduced_train_loss
5k
10k
15k
20k
25k
global_step
0.4
0.5
0.6
0.7
0.8
0.9
1
Run: crossbmg4egc_lhmain3cross_oci_FC-GPT_llama_tiny_canaryset_b6s4kf-ASR-AST_lr1e-4wd1e-3_CosineAnnealing_warmup2000_minlr1e-6_gbs1024_mbs16_ep200
445
Run set 2
Run set 5
™
Add a comment