Update `README.md`

README.md

@@ -13,13 +13,12 @@ library_name: transformers
 
 In this repo, we present a permissively licensed open source reproduction of Meta AI's [LLaMA](https://ai.facebook.com/blog/large-language-model-llama-meta-ai/) large language model. We are releasing a series of 3B, 7B and 13B models trained on 1T tokens. We provide PyTorch and JAX weights of pre-trained OpenLLaMA models, as well as evaluation results and comparison against the original LLaMA models. The v2 model is better than the old v1 model trained on a different data mixture. Please see the [project homepage of OpenLLaMA](https://github.com/openlm-research/open_llama) for more details.
 
-
-
 ## Weights Release, License and Usage
 
 We release the weights in two formats: an EasyLM format to be use with our [EasyLM framework](https://github.com/young-geng/EasyLM), and a PyTorch format to be used with the [Hugging Face transformers](https://huggingface.co/docs/transformers/index) library. Both our training framework EasyLM and the checkpoint weights are licensed permissively under the Apache 2.0 license.
 
 ### Loading the Weights with Hugging Face Transformers
+
 Preview checkpoints can be directly loaded from Hugging Face Hub. **Please note that it is advised to avoid using the Hugging Face fast tokenizer for now, as we’ve observed that** [**the auto-converted fast tokenizer sometimes gives incorrect tokenizations**](https://github.com/huggingface/transformers/issues/24233)**.** This can be achieved by directly using the `LlamaTokenizer` class, or passing in the `use_fast=False` option for the `AutoTokenizer` class. See the following example for usage.
 
 ```python
@@ -41,16 +40,19 @@ model = LlamaForCausalLM.from_pretrained(
 
 prompt = 'Q: What is the largest animal?\nA:'
 input_ids = tokenizer(prompt, return_tensors="pt").input_ids
+input_ids = input_ids.to(model.device)
 
-generation_output = model.generate(
-    input_ids=input_ids, max_new_tokens=32
-)
-print(tokenizer.decode(generation_output[0]))
+with torch.cuda.amp.autocast():
+    generation_output = model.generate(
+        input_ids=input_ids, max_new_tokens=32
+    )
+    print(tokenizer.decode(generation_output[0]))
 ```
 
 For more advanced usage, please follow the [transformers LLaMA documentation](https://huggingface.co/docs/transformers/main/model_doc/llama).
 
 ### Evaluating with LM-Eval-Harness
+
 The model can be evaluated with [lm-eval-harness](https://github.com/EleutherAI/lm-evaluation-harness). However, due to the aforementioned tokenizer issue, we need to avoid using the fast tokenizer to obtain the correct results. This can be achieved by passing in `use_fast=False` to [this part of lm-eval-harness](https://github.com/EleutherAI/lm-evaluation-harness/blob/4b701e228768052cfae9043dca13e82052ca5eea/lm_eval/models/huggingface.py#LL313C9-L316C10), as shown in the example below:
 
 ```python
@@ -62,9 +64,8 @@ tokenizer = self.AUTO_TOKENIZER_CLASS.from_pretrained(
 ```
 
 ### Loading the Weights with EasyLM
-For using the weights in our EasyLM framework, please refer to the [LLaMA documentation of EasyLM](https://github.com/young-geng/EasyLM/blob/main/docs/llama.md). Note that unlike the original LLaMA model, our OpenLLaMA tokenizer and weights are trained completely from scratch so it is no longer needed to obtain the original LLaMA tokenizer and weights.
-
 
+For using the weights in our EasyLM framework, please refer to the [LLaMA documentation of EasyLM](https://github.com/young-geng/EasyLM/blob/main/docs/llama.md). Note that unlike the original LLaMA model, our OpenLLaMA tokenizer and weights are trained completely from scratch so it is no longer needed to obtain the original LLaMA tokenizer and weights.
 
 ## Dataset and Training
 
@@ -72,14 +73,12 @@ The v1 models are trained on the [RedPajama dataset](https://huggingface.co/data
 
 We train the models on cloud TPU-v4s using [EasyLM](https://github.com/young-geng/EasyLM), a JAX based training pipeline we developed for training and fine-tuning large language models. We employ a combination of normal data parallelism and [fully sharded data parallelism (also know as ZeRO stage 3)](https://engineering.fb.com/2021/07/15/open-source/fsdp/) to balance the training throughput and memory usage. Overall we reach a throughput of over 2200 tokens / second / TPU-v4 chip for our 7B model.
 
-
 ## Evaluation
 
 We evaluated OpenLLaMA on a wide range of tasks using [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness).  The LLaMA results are generated by running the original LLaMA model on the same evaluation metrics. We note that our results for the LLaMA model differ slightly from the original LLaMA paper, which we believe is a result of different evaluation protocols. Similar differences have been reported in [this issue of lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/issues/443). Additionally, we present the results of GPT-J, a 6B parameter model trained on the [Pile](https://pile.eleuther.ai/) dataset by [EleutherAI](https://www.eleuther.ai/).
 
 The original LLaMA model was trained for 1 trillion tokens and GPT-J was trained for 500 billion tokens.  We present the results in the table below. OpenLLaMA exhibits comparable performance to the original LLaMA and GPT-J across a majority of tasks, and outperforms them in some tasks.
 
-
 | **Task/Metric**        | GPT-J 6B | LLaMA 7B | LLaMA 13B | OpenLLaMA 7Bv2 | OpenLLaMA 3B | OpenLLaMA 7B | OpenLLaMA 13B |
 | ---------------------- | -------- | -------- | --------- | -------------- | ------------ | ------------ | ------------- |
 | anli_r1/acc            | 0.32     | 0.35     | 0.35      | 0.34           | 0.33         | 0.33         | 0.33          |
@@ -105,10 +104,8 @@ The original LLaMA model was trained for 1 trillion tokens and GPT-J was trained
 | winogrande/acc         | 0.64     | 0.68     | 0.70      | 0.66           | 0.62         | 0.67         | 0.70          |
 | Average                | 0.52     | 0.55     | 0.57      | 0.56           | 0.53         | 0.55         | 0.57          |
 
-
 We removed the task CB and WSC from our benchmark, as our model performs suspiciously high on these two tasks. We hypothesize that there could be a benchmark data contamination in the training set.
 
-
 ## Contact
 
 We would love to get feedback from the community. If you have any questions, please open an issue or contact us.
@@ -117,15 +114,12 @@ OpenLLaMA is developed by:
 [Xinyang Geng](https://young-geng.xyz/)* and [Hao Liu](https://www.haoliu.site/)* from Berkeley AI Research.
 *Equal Contribution
 
-
-
 ## Acknowledgment
 
 We thank the [Google TPU Research Cloud](https://sites.research.google/trc/about/) program for providing part of the computation resources. We’d like to specially thank Jonathan Caton from TPU Research Cloud for helping us organizing compute resources, Rafi Witten from the Google Cloud team and James Bradbury from the Google JAX team for helping us optimizing our training throughput. We’d also want to thank Charlie Snell, Gautier Izacard, Eric Wallace, Lianmin Zheng and our user community for the discussions and feedback.
 
 The OpenLLaMA 13B v1 model is trained in collaboration with [Stability AI](https://stability.ai/), and we thank Stability AI for providing the computation resources. We’d like to especially thank David Ha and Shivanshu Purohit for the coordinating the logistics and providing engineering support.
 
-
 ## Reference
 
 If you found OpenLLaMA useful in your research or applications, please cite using the following BibTeX:
@@ -155,8 +149,3 @@ If you found OpenLLaMA useful in your research or applications, please cite usin
   year={2023}
 }
 ```
-
-
-
-
-