Per-layer parameter and MAC table
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- Half a day
- Newbie friendliness
- 55/100
- Issue type
- Feature
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- python
- Domain
- machine-learning
Research direction
Start in src/complexity.py, where the forward hooks for Conv2d, BatchNorm2d and Linear modules belong, and check how count_params already totals parameters. Add layer_table(model) and save_layer_table(model, path), applying the MAC formulas from the issue. Done means results/arch_A.csv and results/arch_B.csv exist and their totals match count_params. Issue #16 is listed as a dependency, so confirm it is merged before you start.
Written by the indexing model from the issue text.
Description
Forward hooks that record params and multiply-accumulates per layer.
Files: src/complexity.py
Tasks
- Hook Conv2d, BatchNorm2d and Linear modules
- Conv MACs = H_out * W_out * C_out * k_h * k_w * C_in / groups; Linear MACs = in * out; BN = 0
- Label convs as Conv / DW conv / PW conv
-
layer_table(model)returns list of dicts;save_layer_table(model, path)writes CSV
Done when
results/arch_A.csv and results/arch_B.csv are produced; totals match count_params.
Depends on
#16
Close with a commit or PR message containing Closes #<this issue>.
- Dominant language
- Jupyter Notebook
- Stars
- 0
- Forks
- 0
- Avg merge
- 1m
- Merged PRs (30d)
- 3
Getting set up
This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from VisionForgeCSE/CNN-For-Image-Classification
-
evaluation
Difficulty 2/5 1-3 hours Newbie friendliness 66/100
-
evaluation good first issue
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
-
evaluation
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
training
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
Optimizer factoryOpentraining
Difficulty 2/5 1-3 hours Newbie friendliness 65/100
All issues in VisionForgeCSE/CNN-For-Image-Classification
Similar issues
-
Triage 🩺
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 70/100
-
[Bug]: iter_prefetch crashes on accelerators without record_stream (MPS)Possibly taken @begcdn claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
vllm-project/llm-compressor#3269 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 60/100
Maintainers usually reply within 2 days
-
Potential Bug
Difficulty 2/5 1-3 hours Newbie friendliness 70/100
Maintainers usually reply within 1 day