Update README.md

README.md

@@ -8,15 +8,23 @@ license: mit
 
 [![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/kuprel/min-dalle/blob/main/min_dalle.ipynb)
 [![Discord](https://img.shields.io/discord/823813159592001537?color=5865F2&logo=discord&logoColor=white)](https://discord.com/channels/823813159592001537/912729332311556136)
-[GitHub](https://github.com/kuprel/min-dalle)
+**[GitHub](https://github.com/kuprel/min-dalle)**
+**[❤️ Sponsor](https://github.com/sponsors/kuprel)**
 
-This is a fast, minimal port of [DALL·E Mega](https://wandb.ai/dalle-mini/dalle-mini/reports/DALL-E-Mega-Training-Journal--VmlldzoxODMxMDI2).  It has been stripped down for inference and converted to PyTorch.  The only third party dependencies are numpy, requests, pillow and torch.
+This is a fast, minimal port of Boris Dayma's [DALL·E Mini](https://github.com/borisdayma/dalle-mini) (with mega weights).  It has been stripped down for inference and converted to PyTorch.  The only third party dependencies are numpy, requests, pillow and torch.
 
 To generate a 4x4 grid of DALL·E Mega images it takes:
 - 89 sec with a T4 in Colab
 - 48 sec with a P100 in Colab
 - 13 sec with an A100 on Replicate
 
+Here's a more detailed breakdown of performance on an A100. Credit to [@technobird22](https://github.com/technobird22) and his [NeoGen](https://github.com/technobird22/NeoGen) discord bot for the graph.
+<br />
+<img src="https://github.com/kuprel/min-dalle/raw/main/performance.png" alt="min-dalle" width="450"/>
+<br />
+
+The flax model and code for converting it to torch can be found [here](https://github.com/kuprel/min-dalle-flax).
+
 ## Install
 
 ```bash
@@ -33,20 +41,23 @@ from min_dalle import MinDalle
 model = MinDalle(
     models_root='./pretrained',
     dtype=torch.float32,
+    device='cuda',
     is_mega=True, 
     is_reusable=True
 )
 ```
 
-The required models will be downloaded to `models_root` if they are not already there.  Set the `dtype` to `torch.float16` to save GPU memory.  If you have an Ampere architecture GPU you can use `torch.bfloat16`.  Once everything has finished initializing, call `generate_image` with some text as many times as you want.  Use a positive `seed` for reproducible results.  Higher values for `log2_supercondition_factor` result in better agreement with the text but a narrower variety of generated images.  Every image token is sampled from the top-$k$ most probable tokens.
+The required models will be downloaded to `models_root` if they are not already there.  Set the `dtype` to `torch.float16` to save GPU memory.  If you have an Ampere architecture GPU you can use `torch.bfloat16`.  Set the `device` to either "cuda" or "cpu".  Once everything has finished initializing, call `generate_image` with some text as many times as you want.  Use a positive `seed` for reproducible results.  Higher values for `supercondition_factor` result in better agreement with the text but a narrower variety of generated images.  Every image token is sampled from the `top_k` most probable tokens.  The largest logit is subtracted from the logits to avoid infs.  The logits are then divided by the `temperature`.  If `is_seamless` is true, the image grid will be tiled in token space not pixel space.
 
 ```python
 image = model.generate_image(
     text='Nuclear explosion broccoli',
     seed=-1,
     grid_size=4,
-    log2_k=6,
-    log2_supercondition_factor=5,
+    is_seamless=False,
+    temperature=1,
+    top_k=256,
+    supercondition_factor=32,
     is_verbose=False
 )
 
@@ -54,7 +65,7 @@ display(image)
 ```
 <img src="https://github.com/kuprel/min-dalle/raw/main/examples/nuclear_broccoli.jpg" alt="min-dalle" width="400"/>
 
-credit: [https://twitter.com/hardmaru/status/1544354119527596034](https://twitter.com/hardmaru/status/1544354119527596034)
+Credit to [@hardmaru](https://twitter.com/hardmaru) for the [example](https://twitter.com/hardmaru/status/1544354119527596034)
 
 
 ### Saving Individual Images
@@ -64,9 +75,11 @@ The images can also be generated as a `FloatTensor` in case you want to process
 images = model.generate_images(
     text='Nuclear explosion broccoli',
     seed=-1,
-    image_count=7,
-    log2_k=6,
-    log2_supercondition_factor=5,
+    grid_size=3,
+    is_seamless=False,
+    temperature=1,
+    top_k=256,
+    supercondition_factor=16,
     is_verbose=False
 )
 ```
@@ -81,18 +94,20 @@ image = Image.fromarray(images[i])
 image.save('image_{}.png'.format(i))
 ```
 
-### Interactive
+### Progressive Outputs
 
-If the model is being used interactively (e.g. in a notebook) `generate_image_stream` can be used to generate a stream of images as the model is decoding.  The detokenizer adds a slight delay for each image.  Setting `log2_mid_count` to 3 results in a total of `2 ** 3 = 8` generated images.  The only valid values for `log2_mid_count` are 0, 1, 2, 3, and 4.  This is implemented in the colab.
+If the model is being used interactively (e.g. in a notebook) `generate_image_stream` can be used to generate a stream of images as the model is decoding.  The detokenizer adds a slight delay for each image.  Set `progressive_outputs` to `True` to enable this.  An example is implemented in the colab.
 
 ```python
 image_stream = model.generate_image_stream(
     text='Dali painting of WALL·E',
     seed=-1,
     grid_size=3,
-    log2_mid_count=3,
-    log2_k=6,
-    log2_supercondition_factor=3,
+    progressive_outputs=True,
+    is_seamless=False,
+    temperature=1,
+    top_k=256,
+    supercondition_factor=16,
     is_verbose=False
 )
 
@@ -108,8 +123,4 @@ Use `image_from_text.py` to generate images from the command line.
 ```bash
 $ python image_from_text.py --text='artificial intelligence' --no-mega
 ```
-<img src="https://github.com/kuprel/min-dalle/raw/main/examples/artificial_intelligence.jpg" alt="min-dalle" width="200"/>
-
-<br />
-
-[Sponsor this work](https://github.com/sponsors/kuprel)
+<img src="https://github.com/kuprel/min-dalle/raw/main/examples/artificial_intelligence.jpg" alt="min-dalle" width="200"/>