Cdpierse's group workspace
Group: baseline
Name
29 visualized
•
Job Type: null
Job Type: null
29
State
Notes
User
Tags
Created
Runtime
Sweep
_n_gpu
_name_or_path
activation_dropout
adafactor
adam_beta1
adam_beta2
adam_epsilon
add_cross_attention
apply_spec_augment
architectures
attention_dropout
bos_token_id
chunk_size_feed_forward
conv_bias
conv_dim
conv_kernel
conv_stride
ctc_loss_reduction
ctc_zero_infinity
dataloader_drop_last
dataloader_num_workers
dataloader_pin_memory
ddp_find_unused_parameters
debug
deepspeed
disable_tqdm
diversity_penalty
do_eval
do_predict
do_sample
do_stable_layer_norm
do_train
early_stopping
encoder_no_repeat_ngram_size
eos_token_id
eval_accumulation_steps
eval_batch_size
eval_steps
evaluation_strategy
feat_extract_activation
feat_extract_dropout
feat_extract_norm
feat_proj_dropout
final_dropout
Crashed
Using the same data pre-processing and Jim O Regan, with Boris hyperparameters
morgan
9d 18h 50m 38s
-
1
["./artifacts/model-37y72qa8:v0","facebook/wav2vec2-large-xlsr-53"]
0.047292
false
0.9
0.999
1.0000e-8
false
true
["Wav2Vec2ForCTC","Wav2Vec2Model"]
0.10408
1
0
true
512
3.71429
2.42857
mean
true
false
13.33333
true
None
false
None
false
0
None
false
false
true
false
false
0
2
None
34
64
steps
gelu
0
layer
0.0325
0.0083333
1-1
of 1