GGUF
English
reach-vb HF staff commited on
Commit
4046b3f
โ€ข
1 Parent(s): 01b83a1

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +66 -0
README.md ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ datasets:
4
+ - cerebras/SlimPajama-627B
5
+ - bigcode/starcoderdata
6
+ - HuggingFaceH4/ultrachat_200k
7
+ - HuggingFaceH4/ultrafeedback_binarized
8
+ language:
9
+ - en
10
+ widget:
11
+ - example_title: Fibonacci (Python)
12
+ messages:
13
+ - role: system
14
+ content: You are a chatbot who can help code!
15
+ - role: user
16
+ content: Write me a function to calculate the first 10 digits of the fibonacci sequence in Python and print it out to the CLI.
17
+ ---
18
+ <div align="center">
19
+
20
+ # TinyLlama-1.1B
21
+ </div>
22
+
23
+ https://github.com/jzhang38/TinyLlama
24
+
25
+ The TinyLlama project aims to **pretrain** a **1.1B Llama model on 3 trillion tokens**. With some proper optimization, we can achieve this within a span of "just" 90 days using 16 A100-40G GPUs ๐Ÿš€๐Ÿš€. The training has started on 2023-09-01.
26
+
27
+
28
+ We adopted exactly the same architecture and tokenizer as Llama 2. This means TinyLlama can be plugged and played in many open-source projects built upon Llama. Besides, TinyLlama is compact with only 1.1B parameters. This compactness allows it to cater to a multitude of applications demanding a restricted computation and memory footprint.
29
+
30
+ #### This Model
31
+ This is the chat model finetuned on top of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T). **We follow [HF's Zephyr](https://huggingface.co/HuggingFaceH4/zephyr-7b-alpha)'s training recipe.** The model was " initially fine-tuned on a variant of the [`UltraChat`](https://huggingface.co/datasets/stingning/ultrachat) dataset, which contains a diverse range of synthetic dialogues generated by ChatGPT.
32
+ We then further aligned the model with [๐Ÿค— TRL's](https://github.com/huggingface/trl) `DPOTrainer` on the [openbmb/UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) dataset, which contain 64k prompts and model completions that are ranked by GPT-4."
33
+
34
+
35
+ #### How to use
36
+ You will need the transformers>=4.34
37
+ Do check the [TinyLlama](https://github.com/jzhang38/TinyLlama) github page for more information.
38
+
39
+ ```python
40
+ # Install transformers from source - only needed for versions <= v4.34
41
+ # pip install git+https://github.com/huggingface/transformers.git
42
+ # pip install accelerate
43
+
44
+ import torch
45
+ from transformers import pipeline
46
+
47
+ pipe = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", torch_dtype=torch.bfloat16, device_map="auto")
48
+
49
+ # We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
50
+ messages = [
51
+ {
52
+ "role": "system",
53
+ "content": "You are a friendly chatbot who always responds in the style of a pirate",
54
+ },
55
+ {"role": "user", "content": "How many helicopters can a human eat in one sitting?"},
56
+ ]
57
+ prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
58
+ outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
59
+ print(outputs[0]["generated_text"])
60
+ # <|system|>
61
+ # You are a friendly chatbot who always responds in the style of a pirate.</s>
62
+ # <|user|>
63
+ # How many helicopters can a human eat in one sitting?</s>
64
+ # <|assistant|>
65
+ # ...
66
+ ```