[4.2] UNet [DP, Attn, GELU]

[gauravkuppa]

Description

Explain why we're running this and what we expect.

Planned Start Date:

Depends on Previous Experiment? Y/N Effects of GELU on UNet training

Train Command

python train.py \
--name vanilla_dp_unet_attn_gelu \
--model unet \
--batch MAX \
--person_inputs densepose agnostic \
--cloth_inputs cloth \
--val_check_interval  0.05 \
--self_attn \
--accumulated_batches 64 / MAX \
--activation gelu

Report Results

To report a result, copy this into a comment below:

# Result Description
<!--- 
For Experiment Number, use "Major.minor.patch", e.g. 1.2.0.
Major.minor should match the [M.m] in the title. 
Patch describes a bug fix (change in the code or branch).
-->
**Experiment Number:** 1.2.0
**Branch:** `master`
**Timestamp:** MM/DD/YYYY 9pm PT
**Epochs:** 

# Architecture
**Model Layers:**
<!-- Paste the printed Model Layers -->

**Module Parameters:**
<!-- Paste the Params table -->

# Loss Graphs
<!--- Put detailed loss graphs here. Please include all graphs! -->

# Image Results
<!--- Put detailed image results here. Please include all images! Multiple screenshots is good. -->

# Comments, Observations, or Insights
<!--- Optional -->

[veralauee]

Result Description

Experiment Number: 4.2 Branch: master Timestamp: 09/12/2020 10am PT Epochs: 1

Module Parameters: batch size: 4

Loss Graphs

Image Results

train

val

[veralauee]

Result Description

Experiment Number: 4.2.0 Branch: master Timestamp: 09/14/2020 6pm PT Epochs: finish

Loss Graphs

Image Results

[andrewjong]

I'm extremely impressed at the validation loss curve for L1 and VGG

[gauravkuppa]

Experiment 4.2.1 Fixed #115

Train Command

python train.py \
--model unet \
--batch 16 \
--person_inputs densepose agnostic \
--cloth_inputs cloth \
--val_check_interval 0.05 \
--self_attn \
--accumulated_batches 4 \
-j 12 \
--activation gelu \
--name 4.2.1 \
--vvt_dataroot ~/data/fw_gan_vvt

Report Results

To report a result, copy this into a comment below:

# Result Description
<!--- 
For Experiment Number, use "Major.minor.patch", e.g. 1.2.0.
Major.minor should match the [M.m] in the title. 
Patch describes a bug fix (change in the code or branch).
-->
**Experiment Number:** 1.2.0
**Branch:** `master`
**Timestamp:** MM/DD/YYYY 9pm PT
**Epochs:** 

# Architecture
**Model Layers:**
<!-- Paste the printed Model Layers -->

**Module Parameters:**
<!-- Paste the Params table -->

# Loss Graphs
<!--- Put detailed loss graphs here. Please include all graphs! -->

# Image Results
<!--- Put detailed image results here. Please include all images! Multiple screenshots is good. -->

# Comments, Observations, or Insights
<!--- Optional -->

• [x] Open GitHub Issue
• [x] Start training with tmux (tensorboard and training)
• [ ] Upload scalars, train, and validation images to GitHub
• [ ] Upload checkpoints to Google Drive
• [ ] Generate test results from latest epoch
• [ ] Calculate metrics (PSNR, SSIM)
• [ ] Visualize metrics