init evaluation, memory, elapsed time

README.md

@@ -1,4 +1,5 @@
 # ⚡ Image to Latent Representations
+
 ### A compression towards Latent Space
 
 <p align='center'>
@@ -7,19 +8,23 @@
 
 [EXPERIMENTAL]
 
-A world of compression without original images, latent representations is all you need. 
+In a world of compression without storing original images, latent space representations are all you need....?
 
 **Stack**
+
 - Models
-  - VQVAE: pretraining (see notebooks)
-  - VAE Tiny: madebyollin/taesd
-  - Stable Diffusion: Lykon/dreamshaper-7
-- Optimization
-  - Attention Slicing
-  - LCM LoRa: latent-consistency/lcm-lora-sdv1-5
-  - Fuse LoRa
+    - VQVAE: pretraining (see notebooks)
+    - VAE Tiny: madebyollin/taesd
+    - Stable Diffusion: Lykon/dreamshaper-7
+- Optimization [Stable Diffusion]
+    - Attention Slicing
+    - LCM LoRa: latent-consistency/lcm-lora-sdv1-5
+    - Fuse LoRa
+- Flavour
+    - 8bit latent space
 
 ### Preview
+
 <p align='center'>
   <img width='50%' src="/asset/astronaut.png">
 </p>
@@ -28,20 +33,42 @@ A world of compression without original images, latent representations is all yo
   <img width='50%' src="/asset/similar_latents.png">
 </p>
 
+### Evaluation
+
+| Memory    | n(X) | Q1 MB  | Q2 MB  | Q3  MB | Σ  MB  |
+|-----------|------|--------|--------|--------|--------|
+| Originals | 99   | 0.299  | 0.338  | 0.376  | 33.631 |
+| Latents   | 99   | 0.0127 | 0.0131 | 0.0134 | 1.294  |
+
+| Elapsed time (ms)                            | µ      | m      | σ     | min    | max     | 1 run    |
+|----------------------------------------------|--------|--------|-------|--------|---------|----------|
+| Originals, file store (1)                    | 1.856  | 1.761  | 0.889 | 1.037  | 8.913   | -        |
+| Originals, file store (99)                   | -      | -      | -     | -      | -       | 195.232  |
+| Latents, file store (1)                      | 1.255  | 1.125  | 0.418 | 0.99   | 3.961   | -        |
+| Latents with reconstruction, file store (1)  | -      | -      | -     | -      | -       | 40.559   |
+| Latents with reconstruction, file store (99) | -      | -      | -     | -      | -       | 2522.841 |
+| Latents, vector database (1)                 | 63.007 | 63.717 | 8.502 | 42.864 | 104.509 | -        |
+| Latents, vector database (99)                | -      | -      | -     | -      | -       | 4189     |
+
 ## 🚀 Prerequisite
+
 - install [miniforge](https://github.com/conda-forge/miniforge)
 - create virtual env || conda
 - initialize [Qdrant](https://qdrant.tech)
 - from root enter the following command line
+
 ```commandline
 pip install -r requirements.txt
 ```
+
 ```commandline
 pip install python-dotenv
 ```
 
 ### **WINDOWS** for CUDA Deep Neural Network
+
 - tensorflow
+
 ```commandline 
 conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
 ```
@@ -51,12 +78,15 @@ pip install tensorflow==2.10
 ```
 
 - [pytorch](https://pytorch.org/get-started)
+
 ```commandline
 pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121
 ```
 
 ### **MACOS** for MPS
+
 - tensorflow [installer](https://developer.apple.com/metal/tensorflow-plugin/) for MPS
+
 ```commandline 
 conda install -c apple tensorflow-deps
 ```
@@ -66,27 +96,30 @@ pip install tensorflow-macos==2.10.0 tensorflow-metal==0.6.0
 ```
 
 - [pytorch](https://pytorch.org/get-started)
+
 ```commandline 
 pip install torch torchvision
 ```
 
 - app
+
 ```commandline
 python app/main.py
 ```
 
 - webapp
+
 ```commandline
 streamlit run webapp.py
 ```
 
-
 ## 📖 DOCS
+
 - http://127.0.0.1:8000/docs
 
 ## ARCHITECTURE
 
-
 # !!Credits
+
 - [madebyollin](https://github.com/madebyollin)
 - 🤗 [Hugging Face](https://github.com/huggingface)