refiners/tests/training_utils/mock_config.toml

[mock_model]
requires_grad = true
use_activation = true

[clock]
verbose = false

[gradient_clipping]
clip_grad_norm = 1.0

[training]
duration = "100:epoch"
seed = 0
device = "cpu"
dtype = "float32"
batch_size = 4
gradient_accumulation = "4:step"
evaluation_interval = "5:epoch"
evaluation_seed = 1

[optimizer]
optimizer = "SGD"
learning_rate = 1

[lr_scheduler]
type = "ConstantLR"
update_interval = "1:step"
warmup = "20:step"
refactor register_model decorator 2024-02-12 13:17:51 +00:00			`[mock_model]`
Allow optional train ModelConfig + forbid extra input for configs 2024-02-10 14:53:18 +00:00			`requires_grad = true`
Enforce correct subtype for the config param in both decorators Also add a custom ModelConfig for the MockTrainer test Update src/refiners/training_utils/config.py Co-authored-by: Cédric Deltheil <355031+deltheil@users.noreply.github.com> 2024-02-12 14:53:24 +00:00			`use_activation = true`
Allow optional train ModelConfig + forbid extra input for configs 2024-02-10 14:53:18 +00:00
add @register_model and @register_callback decorators Refactor ClockTrainer to include Callback 2024-02-12 08:28:41 +00:00			`[clock]`
			`verbose = false`

			`[gradient_clipping]`
			`clip_grad_norm = 1.0`
add basic unit test for training_utils 2024-01-14 14:06:48 +00:00
			`[training]`
			`duration = "100:epoch"`
			`seed = 0`
make device and dtype work in Trainer class 2024-02-06 21:39:38 +00:00			`device = "cpu"`
			`dtype = "float32"`
add basic unit test for training_utils 2024-01-14 14:06:48 +00:00			`batch_size = 4`
			`gradient_accumulation = "4:step"`
			`evaluation_interval = "5:epoch"`
			`evaluation_seed = 1`

			`[optimizer]`
			`optimizer = "SGD"`
			`learning_rate = 1`

rename Scheduler -> LRScheduler 2024-02-15 09:48:12 +00:00			`[lr_scheduler]`
			`type = "ConstantLR"`
add basic unit test for training_utils 2024-01-14 14:06:48 +00:00			`update_interval = "1:step"`
			`warmup = "20:step"`