Corrected name of package "cuda12-pip" (#194)

The `cuda12-pip` package was wrongly named `cuda12_pip`
in requirements.txt

README.md

@@ -4,27 +4,21 @@ This repository contains JAX example code for loading and running the Grok-1 ope
 
 Make sure to download the checkpoint and place the `ckpt-0` directory in `checkpoints` - see [Downloading the weights](#downloading-the-weights)
 
-## 1. Installation
+Then, run
 
-1. Install the project dependencies
+```shell
-
-```bash
 pip install -r requirements.txt
-```
-
-2. Run the project
-
-```bash
 python run.py
 ```
 
+to test the code.
+
 The script loads the checkpoint and samples from the model on a test input.
 
-Due to the large size of the model (314 Billion parameters), a machine with enough GPU memory is required to test the model with the example code.
+Due to the large size of the model (314B parameters), a machine with enough GPU memory is required to test the model with the example code.
-
 The implementation of the MoE layer in this repository is not efficient. The implementation was chosen to avoid the need for custom kernels to validate the correctness of the model.
 
-## 2. Model Specifications
+# Model Specifications
 
 Grok-1 is currently designed with the following specifications:
 
@@ -39,9 +33,8 @@ Grok-1 is currently designed with the following specifications:
   - Rotary embeddings (RoPE)
   - Supports activation sharding and 8-bit quantization
 - **Maximum Sequence Length (context):** 8,192 tokens
-- **TPU/GPU:** NVIDIA/AMD supported only
 
-## 3. Downloading the weights
+# Downloading the weights
 
 You can download the weights using a torrent client and this magnet link:
 

requirements.txt

@@ -1,4 +1,4 @@
-dm-haiku==0.0.12
+dm_haiku==0.0.12
 jax[cuda12-pip]==0.4.25 -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
 numpy==1.26.4
 sentencepiece==0.2.0