Fix: Typo.

README.md

@@ -2,8 +2,8 @@
 license: other
 ---
 
-# xLSTM goes 7B
-This xLSTM was pre-trained on the DCLM and selected high-quality data for in a total of approx. 2.3 T tokens using the `xlstm-jax` framework.
+# xLSTM-7B
+This xLSTM-7B was pre-trained on the DCLM and selected high-quality data for in a total of approx. 2.3 T tokens using the `xlstm-jax` framework.
 
 
 ## How to use it
@@ -26,13 +26,13 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 xlstm = AutoModelForCausalLM.from_pretrained("NX-AI/xLSTM-7b", device_map="auto")
 
 # this is a fork of EleutherAI/gpt-neox-20b
-tokenizers = AutoTokenizer.from_pretrained("NX-AI/xLSTM-7b")
+tokenizer = AutoTokenizer.from_pretrained("NX-AI/xLSTM-7b")
 
 xlstm(tokenizer("Hello xLSTM, how are you doing?"))
 ```
 
 ## Speed results
-Generation Speed using `torch.cuda.graph` and `torch.compile` optimizations:
+Generation Speed using `torch.cuda.graph` and `torch.compile` optimizations on one NVIDIA H100:
 ![generation speed](plot_tokens_per_sec.svg)
 
 ## Performance
@@ -52,4 +52,3 @@ Using HuggingFace's `lighteval` in the Leaderboard-v1 settings:
 
 ## License 
 NXAI Community License (see `LICENSE` file)
-