Document feedback

Question document fragment

When a question document fragment contains a formula, it is displayed as a space.

Submission type

issue

It's a little complicated...

I'd like to ask someone.

PR

Just a small problem.

I can fix it online!

Please select the submission type

Problem type

Specifications and Common Mistakes

- Specifications and Common Mistakes:

- Misspellings or punctuation mistakes,incorrect formulas, abnormal display.

- Incorrect links, empty cells, or wrong formats.

- Chinese characters in English context.

- Minor inconsistencies between the UI and descriptions.

- Low writing fluency that does not affect understanding.

- Incorrect version numbers, including software package names and version numbers on the UI.

Usability

- Usability:

- Incorrect or missing key steps.

- Missing main function descriptions, keyword explanation, necessary prerequisites, or precautions.

- Ambiguous descriptions, unclear reference, or contradictory context.

- Unclear logic, such as missing classifications, items, and steps.

Correctness

- Correctness:

- Technical principles, function descriptions, supported platforms, parameter types, or exceptions inconsistent with that of software implementation.

- Incorrect schematic or architecture diagrams.

- Incorrect commands or command parameters.

- Incorrect code.

- Commands inconsistent with the functions.

- Wrong screenshots.

- Sample code running error, or running results inconsistent with the expectation.

Risk Warnings

- Risk Warnings:

- Lack of risk warnings for operations that may damage the system or important data.

Content Compliance

- Content Compliance:

- Contents that may violate applicable laws and regulations or geo-cultural context-sensitive words and expressions.

- Copyright infringement.

Please select the type of question

Problem description

Describe the bug so that we can quickly locate the problem.

Submit

Document feedback

mindformers.core.LearningRateWiseLayer

class mindformers.core.LearningRateWiseLayer(base_lr, lr_scale)

Learning Rate Wise Layer.

This approach allows each layer to adapt its learning rate according to its specific needs, leading to more efficient and effective training. The learning rate for each layer is determined by a base learning rate modulated by a scaling factor specific to that layer.

Initially, the learning rate for each layer is set based on a linear scaling strategy:

$${\eta }_{t,l}={\eta }_{\text{base}}\times {\alpha }_l$$

where ${\eta }_{t,l}$ is the learning rate for layer $l$ at time $t$ , ${\eta }_{\text{base}}$ is the base learning rate, and ${\alpha }_l$ is the scaling factor for layer $l$ .

As training progresses, the learning rate for each layer is adjusted according to the following cosine annealing schedule:

$${\eta }_{t,l}={\eta }_{\text{end}}+\frac{1}{2}({\eta }_{\text{base}}\times {\alpha }_l-{\eta }_{\text{end}})(1+\cos \left(\frac{T_{cur}}{T_{max}}\pi \right))$$

where $T_{cur}$ is the number of epochs completed since the learning rate was last reset, and $T_{max}$ is the total number of epochs before the next reset. ${\eta }_{\text{end}}$ represents the minimum learning rate at the end of the training.

Parameters

• base_lr (mindspore.nn.learning_rate_schedule.LearningRateSchedule) – The base learning rate schedule.

• lr_scale (float) – The value for learning rate scaling.

Inputs:

• global_step (int) - The global step.

Outputs:

Learning rate.

Examples

>>> import mindspore as ms
>>> from mindformers.core import LinearWithWarmUpLR
>>> from mindformers.core import LearningRateWiseLayer
>>>
>>> ms.set_context(mode=ms.GRAPH_MODE)
>>> total_steps = 20
>>> warmup_steps = 10
>>> learning_rate = 0.005
>>>
>>> linear_warmup = LinearWithWarmUpLR(learning_rate=learning_rate,
...                                    warmup_steps=warmup_steps,
...                                    total_steps=total_steps)
>>> learning_rate_wise_layer = LearningRateWiseLayer(linear_warmup, 0.5)
>>> print(learning_rate_wise_layer(1))
0.00025
>>> print(learning_rate_wise_layer(15))
0.00125