LLukas22 commited on
Commit
5e17023
1 Parent(s): 9ef37f5

Generated README.md

Browse files
Files changed (1) hide show
  1. README.md +79 -0
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ tags:
5
+ - llm-rs
6
+ - ggml
7
+ pipeline_tag: text-generation
8
+ datasets:
9
+ - the_pile
10
+ ---
11
+
12
+ # GGML covnerted Models of [EleutherAI](https://huggingface.co/EleutherAI)'s GPT-J model
13
+
14
+ ## Description
15
+
16
+ GPT-J 6B is a transformer model trained using Ben Wang's [Mesh Transformer JAX](https://github.com/kingoflolz/mesh-transformer-jax/). "GPT-J" refers to the class of model, while "6B" represents the number of trainable parameters.
17
+
18
+ <figure>
19
+
20
+ | Hyperparameter | Value |
21
+ |----------------------|------------|
22
+ | \\(n_{parameters}\\) | 6053381344 |
23
+ | \\(n_{layers}\\) | 28&ast; |
24
+ | \\(d_{model}\\) | 4096 |
25
+ | \\(d_{ff}\\) | 16384 |
26
+ | \\(n_{heads}\\) | 16 |
27
+ | \\(d_{head}\\) | 256 |
28
+ | \\(n_{ctx}\\) | 2048 |
29
+ | \\(n_{vocab}\\) | 50257/50400&dagger; (same tokenizer as GPT-2/3) |
30
+ | Positional Encoding | [Rotary Position Embedding (RoPE)](https://arxiv.org/abs/2104.09864) |
31
+ | RoPE Dimensions | [64](https://github.com/kingoflolz/mesh-transformer-jax/blob/f2aa66e0925de6593dcbb70e72399b97b4130482/mesh_transformer/layers.py#L223) |
32
+ <figcaption><p><strong>&ast;</strong> Each layer consists of one feedforward block and one self attention block.</p>
33
+ <p><strong>&dagger;</strong> Although the embedding matrix has a size of 50400, only 50257 entries are used by the GPT-2 tokenizer.</p></figcaption></figure>
34
+
35
+ The model consists of 28 layers with a model dimension of 4096, and a feedforward dimension of 16384. The model
36
+ dimension is split into 16 heads, each with a dimension of 256. Rotary Position Embedding (RoPE) is applied to 64
37
+ dimensions of each head. The model is trained with a tokenization vocabulary of 50257, using the same set of BPEs as
38
+ GPT-2/GPT-3.
39
+
40
+
41
+ ## Converted Models
42
+ | Name | Based on | Type | Container | GGML Version |
43
+ |:---------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------|:-------|:------------|:---------------|
44
+ | [gpt-j-6b-f16.bin](https://huggingface.co/rustformers/gpt-j-ggml/blob/main/gpt-j-6b-f16.bin) | [EleutherAI/gpt-j-6b](https://huggingface.co/EleutherAI/gpt-j-6b) | F16 | GGML | V3 |
45
+ | [gpt-j-6b-q4_0.bin](https://huggingface.co/rustformers/gpt-j-ggml/blob/main/gpt-j-6b-q4_0.bin) | [EleutherAI/gpt-j-6b](https://huggingface.co/EleutherAI/gpt-j-6b) | Q4_0 | GGML | V3 |
46
+ | [gpt-j-6b-q4_0-ggjt.bin](https://huggingface.co/rustformers/gpt-j-ggml/blob/main/gpt-j-6b-q4_0-ggjt.bin) | [EleutherAI/gpt-j-6b](https://huggingface.co/EleutherAI/gpt-j-6b) | Q4_0 | GGJT | V3 |
47
+ | [gpt-j-6b-q5_1-ggjt.bin](https://huggingface.co/rustformers/gpt-j-ggml/blob/main/gpt-j-6b-q5_1-ggjt.bin) | [EleutherAI/gpt-j-6b](https://huggingface.co/EleutherAI/gpt-j-6b) | Q5_1 | GGJT | V3 |
48
+
49
+ ## Usage
50
+
51
+ ### Python via [llm-rs](https://github.com/LLukas22/llm-rs-python):
52
+
53
+ #### Installation
54
+ Via pip: `pip install llm-rs`
55
+
56
+ #### Run inference
57
+ ```python
58
+ from llm_rs import AutoModel
59
+
60
+ #Load the model, define any model you like from the list above as the `model_file`
61
+ model = AutoModel.from_pretrained("rustformers/gpt-j-ggml",model_file="gpt-j-6b-q4_0-ggjt.bin")
62
+
63
+ #Generate
64
+ print(model.generate("The meaning of life is"))
65
+ ```
66
+
67
+ ### Rust via [Rustformers/llm](https://github.com/rustformers/llm):
68
+
69
+ #### Installation
70
+ ```
71
+ git clone --recurse-submodules https://github.com/rustformers/llm.git
72
+ cd llm
73
+ cargo build --release
74
+ ```
75
+
76
+ #### Run inference
77
+ ```
78
+ cargo run --release -- gptj infer -m path/to/model.bin -p "Tell me how cool the Rust programming language is:"
79
+ ```