Upload README.md with huggingface_hub

README.md

@@ -1,20 +1,3 @@
----
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: tokens
-    list: int64
-  splits:
-  - name: train
-    num_bytes: 409800
-    num_examples: 50
-  download_size: 109371
-  dataset_size: 409800
----
 # Dataset Card for eoinf/tokenized_dataset_test3
 
 ## Original dataset
@@ -25,13 +8,13 @@ Original dataset: monology/pile-uncopyrighted
 - **Total Tokens**: 51,200
 - **Total Sequences**: 50
 - **Context Length**: 1024 tokens
-- **Tokenizer**: EleutherAI/gpt-neox-20b
+- **Tokenizer**: meta-llama/Llama-2-7b-hf
 - **Format**: Each example contains a single field `tokens` with a list of 1024 token IDs
 
 ## Preprocessing
 
 Each document was:
-1. Tokenized using the EleutherAI/gpt-neox-20b tokenizer
+1. Tokenized using the meta-llama/Llama-2-7b-hf tokenizer
 2. Prefixed with a BOS (beginning of sequence) token
 3. Suffixed with an EOS (end of sequence) token
 4. Packed into fixed-length sequences of 1024 tokens