Nintorac Dev
commited on
Commit
Β·
4f0300d
1
Parent(s):
a36105f
restructure
Browse files- s4-dx7-vc-fir/s4-dx7-vc-fir-00/.gitattributes +2 -0
- s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/config.yaml +168 -0
- s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/hydra.yaml +174 -0
- s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/overrides.yaml +11 -0
- s4-dx7-vc-fir/s4-dx7-vc-fir-00/checkpoints/last.ckpt +3 -0
- s4-dx7-vc-fir/{00.ckpt β s4-dx7-vc-fir-00/checkpoints/val/loss.ckpt} +0 -0
- s4-dx7-vc-fir/s4-dx7-vc-fir-00/config_tree.txt +181 -0
s4-dx7-vc-fir/s4-dx7-vc-fir-00/.gitattributes
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
checkpoints/last.ckpt filter=lfs diff=lfs merge=lfs -text
|
2 |
+
checkpoints/val/loss.ckpt filter=lfs diff=lfs merge=lfs -text
|
s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/config.yaml
ADDED
@@ -0,0 +1,168 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
train:
|
2 |
+
seed: 0
|
3 |
+
name: null
|
4 |
+
interval: epoch
|
5 |
+
monitor: val/loss
|
6 |
+
mode: min
|
7 |
+
ema: 0.0
|
8 |
+
test: false
|
9 |
+
debug: false
|
10 |
+
ignore_warnings: false
|
11 |
+
state:
|
12 |
+
mode: null
|
13 |
+
n_context: 0
|
14 |
+
n_context_eval: ${.n_context}
|
15 |
+
ckpt: null
|
16 |
+
optimizer_param_grouping:
|
17 |
+
bias_weight_decay: false
|
18 |
+
normalization_weight_decay: false
|
19 |
+
disable_dataset: false
|
20 |
+
validate_at_start: false
|
21 |
+
pretrained_model_path: null
|
22 |
+
pretrained_model_strict_load: true
|
23 |
+
pretrained_model_state_hook:
|
24 |
+
_name_: null
|
25 |
+
post_init_hook:
|
26 |
+
_name_: null
|
27 |
+
layer_decay:
|
28 |
+
_name_: null
|
29 |
+
decay: 0.7
|
30 |
+
track_grad_norms: false
|
31 |
+
tolerance:
|
32 |
+
logdir: ./resume
|
33 |
+
id: 420
|
34 |
+
wandb:
|
35 |
+
project: hippo
|
36 |
+
group: ''
|
37 |
+
job_type: training
|
38 |
+
mode: online
|
39 |
+
save_dir: .
|
40 |
+
id: null
|
41 |
+
callbacks:
|
42 |
+
learning_rate_monitor:
|
43 |
+
logging_interval: ${train.interval}
|
44 |
+
timer:
|
45 |
+
step: true
|
46 |
+
inter_step: false
|
47 |
+
epoch: true
|
48 |
+
val: true
|
49 |
+
params:
|
50 |
+
total: true
|
51 |
+
trainable: true
|
52 |
+
fixed: true
|
53 |
+
model_checkpoint:
|
54 |
+
monitor: ${train.monitor}
|
55 |
+
mode: ${train.mode}
|
56 |
+
save_top_k: 1
|
57 |
+
save_last: true
|
58 |
+
dirpath: checkpoints/
|
59 |
+
filename: ${train.monitor}
|
60 |
+
auto_insert_metric_name: false
|
61 |
+
verbose: true
|
62 |
+
rich_model_summary:
|
63 |
+
max_depth: 1
|
64 |
+
rich_progress_bar:
|
65 |
+
refresh_rate: 1
|
66 |
+
leave: true
|
67 |
+
trainer:
|
68 |
+
accelerator: gpu
|
69 |
+
strategy: auto
|
70 |
+
devices: 1
|
71 |
+
accumulate_grad_batches: 2
|
72 |
+
max_epochs: 1000
|
73 |
+
gradient_clip_val: null
|
74 |
+
log_every_n_steps: 10
|
75 |
+
limit_train_batches: 1000
|
76 |
+
limit_val_batches: 100
|
77 |
+
enable_model_summary: false
|
78 |
+
loader:
|
79 |
+
batch_size: 14
|
80 |
+
num_workers: 4
|
81 |
+
pin_memory: true
|
82 |
+
drop_last: true
|
83 |
+
dataset:
|
84 |
+
_name_: dx7
|
85 |
+
bits: 8
|
86 |
+
__l_max: 16000
|
87 |
+
sr: 8000
|
88 |
+
duration: 2.5
|
89 |
+
limit: 20000
|
90 |
+
bit_rate: 8
|
91 |
+
task:
|
92 |
+
_name_: base
|
93 |
+
loss: cross_entropy
|
94 |
+
metrics:
|
95 |
+
- bpb
|
96 |
+
- accuracy
|
97 |
+
- accuracy@3
|
98 |
+
- accuracy@5
|
99 |
+
- accuracy@10
|
100 |
+
torchmetrics: null
|
101 |
+
optimizer:
|
102 |
+
_name_: adamw
|
103 |
+
lr: 0.001
|
104 |
+
weight_decay: 0.0
|
105 |
+
betas:
|
106 |
+
- 0.9
|
107 |
+
- 0.999
|
108 |
+
scheduler:
|
109 |
+
_name_: plateau
|
110 |
+
mode: ${train.mode}
|
111 |
+
factor: 0.2
|
112 |
+
patience: 20
|
113 |
+
min_lr: 0.0
|
114 |
+
model:
|
115 |
+
layer:
|
116 |
+
_name_: s4
|
117 |
+
d_state: 64
|
118 |
+
channels: 1
|
119 |
+
bidirectional: false
|
120 |
+
gate: null
|
121 |
+
gate_act: id
|
122 |
+
bottleneck: null
|
123 |
+
activation: gelu
|
124 |
+
mult_act: null
|
125 |
+
final_act: glu
|
126 |
+
postact: null
|
127 |
+
initializer: null
|
128 |
+
weight_norm: false
|
129 |
+
tie_dropout: ${oc.select:model.tie_dropout,null}
|
130 |
+
layer: fftconv
|
131 |
+
mode: nplr
|
132 |
+
init: legs
|
133 |
+
measure: null
|
134 |
+
rank: 1
|
135 |
+
dt_min: 0.001
|
136 |
+
dt_max: 0.1
|
137 |
+
dt_transform: softplus
|
138 |
+
lr:
|
139 |
+
dt: 0.001
|
140 |
+
A: 0.001
|
141 |
+
B: 0.001
|
142 |
+
wd: 0.0
|
143 |
+
n_ssm: 1
|
144 |
+
drop_kernel: 0.0
|
145 |
+
deterministic: false
|
146 |
+
l_max: null
|
147 |
+
verbose: true
|
148 |
+
_name_: sashimi
|
149 |
+
d_model: 64
|
150 |
+
n_layers: 8
|
151 |
+
pool:
|
152 |
+
- 4
|
153 |
+
- 4
|
154 |
+
expand: 2
|
155 |
+
ff: 2
|
156 |
+
prenorm: true
|
157 |
+
dropout: 0.0
|
158 |
+
dropres: 0.0
|
159 |
+
initializer: null
|
160 |
+
transposed: true
|
161 |
+
residual: R
|
162 |
+
norm: layer
|
163 |
+
interp: 0
|
164 |
+
act_pool: null
|
165 |
+
encoder: embedding
|
166 |
+
decoder:
|
167 |
+
_name_: sequence
|
168 |
+
mode: last
|
s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/hydra.yaml
ADDED
@@ -0,0 +1,174 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
hydra:
|
2 |
+
run:
|
3 |
+
dir: ./outputs/${now:%Y-%m-%d}/${now:%H-%M-%S-%f}
|
4 |
+
sweep:
|
5 |
+
dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
|
6 |
+
subdir: ${hydra.job.num}
|
7 |
+
launcher:
|
8 |
+
_target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
|
9 |
+
sweeper:
|
10 |
+
_target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
|
11 |
+
max_batch_size: null
|
12 |
+
params: null
|
13 |
+
help:
|
14 |
+
app_name: ${hydra.job.name}
|
15 |
+
header: '${hydra.help.app_name} is powered by Hydra.
|
16 |
+
|
17 |
+
'
|
18 |
+
footer: 'Powered by Hydra (https://hydra.cc)
|
19 |
+
|
20 |
+
Use --hydra-help to view Hydra specific help
|
21 |
+
|
22 |
+
'
|
23 |
+
template: '${hydra.help.header}
|
24 |
+
|
25 |
+
== Configuration groups ==
|
26 |
+
|
27 |
+
Compose your configuration from those groups (group=option)
|
28 |
+
|
29 |
+
|
30 |
+
$APP_CONFIG_GROUPS
|
31 |
+
|
32 |
+
|
33 |
+
== Config ==
|
34 |
+
|
35 |
+
Override anything in the config (foo.bar=value)
|
36 |
+
|
37 |
+
|
38 |
+
$CONFIG
|
39 |
+
|
40 |
+
|
41 |
+
${hydra.help.footer}
|
42 |
+
|
43 |
+
'
|
44 |
+
hydra_help:
|
45 |
+
template: 'Hydra (${hydra.runtime.version})
|
46 |
+
|
47 |
+
See https://hydra.cc for more info.
|
48 |
+
|
49 |
+
|
50 |
+
== Flags ==
|
51 |
+
|
52 |
+
$FLAGS_HELP
|
53 |
+
|
54 |
+
|
55 |
+
== Configuration groups ==
|
56 |
+
|
57 |
+
Compose your configuration from those groups (For example, append hydra/job_logging=disabled
|
58 |
+
to command line)
|
59 |
+
|
60 |
+
|
61 |
+
$HYDRA_CONFIG_GROUPS
|
62 |
+
|
63 |
+
|
64 |
+
Use ''--cfg hydra'' to Show the Hydra config.
|
65 |
+
|
66 |
+
'
|
67 |
+
hydra_help: ???
|
68 |
+
hydra_logging:
|
69 |
+
version: 1
|
70 |
+
formatters:
|
71 |
+
simple:
|
72 |
+
format: '[%(asctime)s][HYDRA] %(message)s'
|
73 |
+
handlers:
|
74 |
+
console:
|
75 |
+
class: logging.StreamHandler
|
76 |
+
formatter: simple
|
77 |
+
stream: ext://sys.stdout
|
78 |
+
root:
|
79 |
+
level: INFO
|
80 |
+
handlers:
|
81 |
+
- console
|
82 |
+
loggers:
|
83 |
+
logging_example:
|
84 |
+
level: DEBUG
|
85 |
+
disable_existing_loggers: false
|
86 |
+
job_logging:
|
87 |
+
version: 1
|
88 |
+
formatters:
|
89 |
+
simple:
|
90 |
+
format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
|
91 |
+
handlers:
|
92 |
+
console:
|
93 |
+
class: logging.StreamHandler
|
94 |
+
formatter: simple
|
95 |
+
stream: ext://sys.stdout
|
96 |
+
file:
|
97 |
+
class: logging.FileHandler
|
98 |
+
formatter: simple
|
99 |
+
filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
|
100 |
+
root:
|
101 |
+
level: INFO
|
102 |
+
handlers:
|
103 |
+
- console
|
104 |
+
- file
|
105 |
+
disable_existing_loggers: false
|
106 |
+
env: {}
|
107 |
+
mode: RUN
|
108 |
+
searchpath: []
|
109 |
+
callbacks: {}
|
110 |
+
output_subdir: .hydra
|
111 |
+
overrides:
|
112 |
+
hydra:
|
113 |
+
- hydra.mode=RUN
|
114 |
+
task:
|
115 |
+
- +experiment=audio/sashimi-sc09
|
116 |
+
- loader.batch_size=14
|
117 |
+
- dataset=dx7
|
118 |
+
- +dataset.sr=8000
|
119 |
+
- +dataset.duration=2.5
|
120 |
+
- trainer.limit_train_batches=1000
|
121 |
+
- trainer.limit_val_batches=100
|
122 |
+
- trainer.accumulate_grad_batches=2
|
123 |
+
- +dataset.limit=20000
|
124 |
+
- +dataset.bit_rate=8
|
125 |
+
- +tolerance.id=420
|
126 |
+
job:
|
127 |
+
name: train
|
128 |
+
chdir: null
|
129 |
+
override_dirname: +dataset.bit_rate=8,+dataset.duration=2.5,+dataset.limit=20000,+dataset.sr=8000,+experiment=audio/sashimi-sc09,+tolerance.id=420,dataset=dx7,loader.batch_size=14,trainer.accumulate_grad_batches=2,trainer.limit_train_batches=1000,trainer.limit_val_batches=100
|
130 |
+
id: ???
|
131 |
+
num: ???
|
132 |
+
config_name: config.yaml
|
133 |
+
env_set: {}
|
134 |
+
env_copy: []
|
135 |
+
config:
|
136 |
+
override_dirname:
|
137 |
+
kv_sep: '='
|
138 |
+
item_sep: ','
|
139 |
+
exclude_keys: []
|
140 |
+
runtime:
|
141 |
+
version: 1.3.2
|
142 |
+
version_base: '1.1'
|
143 |
+
cwd: /home/ubuntu/s4/s4_dx7/s4
|
144 |
+
config_sources:
|
145 |
+
- path: hydra.conf
|
146 |
+
schema: pkg
|
147 |
+
provider: hydra
|
148 |
+
- path: /home/ubuntu/s4/s4_dx7/s4/configs
|
149 |
+
schema: file
|
150 |
+
provider: main
|
151 |
+
- path: ''
|
152 |
+
schema: structured
|
153 |
+
provider: schema
|
154 |
+
output_dir: /home/ubuntu/s4/s4_dx7/s4/outputs/2024-01-20/12-12-19-068887
|
155 |
+
choices:
|
156 |
+
experiment: audio/sashimi-sc09
|
157 |
+
model: sashimi
|
158 |
+
model/layer: s4
|
159 |
+
scheduler: plateau
|
160 |
+
optimizer: adamw
|
161 |
+
task: multiclass_classification
|
162 |
+
dataset: dx7
|
163 |
+
loader: default
|
164 |
+
trainer: default
|
165 |
+
hydra/env: default
|
166 |
+
hydra/callbacks: null
|
167 |
+
hydra/job_logging: default
|
168 |
+
hydra/hydra_logging: default
|
169 |
+
hydra/hydra_help: default
|
170 |
+
hydra/help: default
|
171 |
+
hydra/sweeper: basic
|
172 |
+
hydra/launcher: basic
|
173 |
+
hydra/output: default
|
174 |
+
verbose: false
|
s4-dx7-vc-fir/s4-dx7-vc-fir-00/.hydra/overrides.yaml
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
- +experiment=audio/sashimi-sc09
|
2 |
+
- loader.batch_size=14
|
3 |
+
- dataset=dx7
|
4 |
+
- +dataset.sr=8000
|
5 |
+
- +dataset.duration=2.5
|
6 |
+
- trainer.limit_train_batches=1000
|
7 |
+
- trainer.limit_val_batches=100
|
8 |
+
- trainer.accumulate_grad_batches=2
|
9 |
+
- +dataset.limit=20000
|
10 |
+
- +dataset.bit_rate=8
|
11 |
+
- +tolerance.id=420
|
s4-dx7-vc-fir/s4-dx7-vc-fir-00/checkpoints/last.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7cf25558fb420f9e29ce11c3198dc8b49785702efde3834fa8f925657806742c
|
3 |
+
size 57590733
|
s4-dx7-vc-fir/{00.ckpt β s4-dx7-vc-fir-00/checkpoints/val/loss.ckpt}
RENAMED
File without changes
|
s4-dx7-vc-fir/s4-dx7-vc-fir-00/config_tree.txt
ADDED
@@ -0,0 +1,181 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
CONFIG
|
2 |
+
βββ train
|
3 |
+
β βββ seed: 0
|
4 |
+
β name: null
|
5 |
+
β interval: epoch
|
6 |
+
β monitor: val/loss
|
7 |
+
β mode: min
|
8 |
+
β ema: 0.0
|
9 |
+
β test: false
|
10 |
+
β debug: false
|
11 |
+
β ignore_warnings: false
|
12 |
+
β state:
|
13 |
+
β mode: null
|
14 |
+
β n_context: 0
|
15 |
+
β n_context_eval: 0
|
16 |
+
β ckpt: null
|
17 |
+
β optimizer_param_grouping:
|
18 |
+
β bias_weight_decay: false
|
19 |
+
β normalization_weight_decay: false
|
20 |
+
β disable_dataset: false
|
21 |
+
β validate_at_start: false
|
22 |
+
β pretrained_model_path: null
|
23 |
+
β pretrained_model_strict_load: true
|
24 |
+
β pretrained_model_state_hook:
|
25 |
+
β _name_: null
|
26 |
+
β post_init_hook:
|
27 |
+
β _name_: null
|
28 |
+
β layer_decay:
|
29 |
+
β _name_: null
|
30 |
+
β decay: 0.7
|
31 |
+
β track_grad_norms: false
|
32 |
+
β
|
33 |
+
βββ tolerance
|
34 |
+
β βββ logdir: ./resume
|
35 |
+
β id: 420
|
36 |
+
β
|
37 |
+
βββ wandb
|
38 |
+
β βββ project: hippo
|
39 |
+
β group: ''
|
40 |
+
β job_type: training
|
41 |
+
β mode: online
|
42 |
+
β save_dir: .
|
43 |
+
β id: null
|
44 |
+
β
|
45 |
+
βββ callbacks
|
46 |
+
β βββ learning_rate_monitor:
|
47 |
+
β logging_interval: epoch
|
48 |
+
β timer:
|
49 |
+
β step: true
|
50 |
+
β inter_step: false
|
51 |
+
β epoch: true
|
52 |
+
β val: true
|
53 |
+
β params:
|
54 |
+
β total: true
|
55 |
+
β trainable: true
|
56 |
+
β fixed: true
|
57 |
+
β model_checkpoint:
|
58 |
+
β monitor: val/loss
|
59 |
+
β mode: min
|
60 |
+
β save_top_k: 1
|
61 |
+
β save_last: true
|
62 |
+
β dirpath: checkpoints/
|
63 |
+
β filename: val/loss
|
64 |
+
β auto_insert_metric_name: false
|
65 |
+
β verbose: true
|
66 |
+
β rich_model_summary:
|
67 |
+
β max_depth: 1
|
68 |
+
β rich_progress_bar:
|
69 |
+
β refresh_rate: 1
|
70 |
+
β leave: true
|
71 |
+
β
|
72 |
+
βββ trainer
|
73 |
+
β βββ accelerator: gpu
|
74 |
+
β strategy: auto
|
75 |
+
β devices: 1
|
76 |
+
β accumulate_grad_batches: 2
|
77 |
+
β max_epochs: 1000
|
78 |
+
β gradient_clip_val: null
|
79 |
+
β log_every_n_steps: 10
|
80 |
+
β limit_train_batches: 1000
|
81 |
+
β limit_val_batches: 100
|
82 |
+
β enable_model_summary: false
|
83 |
+
β
|
84 |
+
βββ loader
|
85 |
+
β βββ batch_size: 14
|
86 |
+
β num_workers: 4
|
87 |
+
β pin_memory: true
|
88 |
+
β drop_last: true
|
89 |
+
β
|
90 |
+
βββ dataset
|
91 |
+
β βββ _name_: dx7
|
92 |
+
β bits: 8
|
93 |
+
β sr: 8000
|
94 |
+
β duration: 2.5
|
95 |
+
β limit: 20000
|
96 |
+
β bit_rate: 8
|
97 |
+
β
|
98 |
+
βββ task
|
99 |
+
β βββ _name_: base
|
100 |
+
β loss: cross_entropy
|
101 |
+
β metrics:
|
102 |
+
β - bpb
|
103 |
+
β - accuracy
|
104 |
+
β - accuracy@3
|
105 |
+
β - accuracy@5
|
106 |
+
β - accuracy@10
|
107 |
+
β torchmetrics: null
|
108 |
+
β
|
109 |
+
βββ optimizer
|
110 |
+
β βββ _name_: adamw
|
111 |
+
β lr: 0.001
|
112 |
+
β weight_decay: 0.0
|
113 |
+
β betas:
|
114 |
+
β - 0.9
|
115 |
+
β - 0.999
|
116 |
+
β
|
117 |
+
βββ scheduler
|
118 |
+
β βββ _name_: plateau
|
119 |
+
β mode: min
|
120 |
+
β factor: 0.2
|
121 |
+
β patience: 20
|
122 |
+
β min_lr: 0.0
|
123 |
+
β
|
124 |
+
βββ model
|
125 |
+
β βββ layer:
|
126 |
+
β _name_: s4
|
127 |
+
β d_state: 64
|
128 |
+
β channels: 1
|
129 |
+
β bidirectional: false
|
130 |
+
β gate: null
|
131 |
+
β gate_act: id
|
132 |
+
β bottleneck: null
|
133 |
+
β activation: gelu
|
134 |
+
β mult_act: null
|
135 |
+
β final_act: glu
|
136 |
+
β postact: null
|
137 |
+
β initializer: null
|
138 |
+
β weight_norm: false
|
139 |
+
β tie_dropout: null
|
140 |
+
β layer: fftconv
|
141 |
+
β mode: nplr
|
142 |
+
β init: legs
|
143 |
+
β measure: null
|
144 |
+
β rank: 1
|
145 |
+
β dt_min: 0.001
|
146 |
+
β dt_max: 0.1
|
147 |
+
β dt_transform: softplus
|
148 |
+
β lr:
|
149 |
+
β dt: 0.001
|
150 |
+
β A: 0.001
|
151 |
+
β B: 0.001
|
152 |
+
β wd: 0.0
|
153 |
+
β n_ssm: 1
|
154 |
+
β drop_kernel: 0.0
|
155 |
+
β deterministic: false
|
156 |
+
β l_max: null
|
157 |
+
β verbose: true
|
158 |
+
β _name_: sashimi
|
159 |
+
β d_model: 64
|
160 |
+
β n_layers: 8
|
161 |
+
β pool:
|
162 |
+
β - 4
|
163 |
+
β - 4
|
164 |
+
β expand: 2
|
165 |
+
β ff: 2
|
166 |
+
β prenorm: true
|
167 |
+
β dropout: 0.0
|
168 |
+
β dropres: 0.0
|
169 |
+
β initializer: null
|
170 |
+
β transposed: true
|
171 |
+
β residual: R
|
172 |
+
β norm: layer
|
173 |
+
β interp: 0
|
174 |
+
β act_pool: null
|
175 |
+
β
|
176 |
+
βββ encoder
|
177 |
+
β βββ embedding
|
178 |
+
βββ decoder
|
179 |
+
βββ _name_: sequence
|
180 |
+
mode: last
|
181 |
+
|