RichardErkhov
commited on
Commit
•
e3d81a2
1
Parent(s):
62bf66b
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Quantization made by Richard Erkhov.
|
2 |
+
|
3 |
+
[Github](https://github.com/RichardErkhov)
|
4 |
+
|
5 |
+
[Discord](https://discord.gg/pvy7H8DZMG)
|
6 |
+
|
7 |
+
[Request more models](https://github.com/RichardErkhov/quant_request)
|
8 |
+
|
9 |
+
|
10 |
+
TinyLlama-1.1B-Tele - GGUF
|
11 |
+
- Model creator: https://huggingface.co/AliMaatouk/
|
12 |
+
- Original model: https://huggingface.co/AliMaatouk/TinyLlama-1.1B-Tele/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [TinyLlama-1.1B-Tele.Q2_K.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q2_K.gguf) | Q2_K | 0.4GB |
|
18 |
+
| [TinyLlama-1.1B-Tele.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.IQ3_XS.gguf) | IQ3_XS | 0.44GB |
|
19 |
+
| [TinyLlama-1.1B-Tele.IQ3_S.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.IQ3_S.gguf) | IQ3_S | 0.47GB |
|
20 |
+
| [TinyLlama-1.1B-Tele.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q3_K_S.gguf) | Q3_K_S | 0.47GB |
|
21 |
+
| [TinyLlama-1.1B-Tele.IQ3_M.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.IQ3_M.gguf) | IQ3_M | 0.48GB |
|
22 |
+
| [TinyLlama-1.1B-Tele.Q3_K.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q3_K.gguf) | Q3_K | 0.51GB |
|
23 |
+
| [TinyLlama-1.1B-Tele.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q3_K_M.gguf) | Q3_K_M | 0.51GB |
|
24 |
+
| [TinyLlama-1.1B-Tele.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q3_K_L.gguf) | Q3_K_L | 0.55GB |
|
25 |
+
| [TinyLlama-1.1B-Tele.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.IQ4_XS.gguf) | IQ4_XS | 0.57GB |
|
26 |
+
| [TinyLlama-1.1B-Tele.Q4_0.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q4_0.gguf) | Q4_0 | 0.59GB |
|
27 |
+
| [TinyLlama-1.1B-Tele.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.IQ4_NL.gguf) | IQ4_NL | 0.6GB |
|
28 |
+
| [TinyLlama-1.1B-Tele.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q4_K_S.gguf) | Q4_K_S | 0.6GB |
|
29 |
+
| [TinyLlama-1.1B-Tele.Q4_K.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q4_K.gguf) | Q4_K | 0.62GB |
|
30 |
+
| [TinyLlama-1.1B-Tele.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q4_K_M.gguf) | Q4_K_M | 0.62GB |
|
31 |
+
| [TinyLlama-1.1B-Tele.Q4_1.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q4_1.gguf) | Q4_1 | 0.65GB |
|
32 |
+
| [TinyLlama-1.1B-Tele.Q5_0.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q5_0.gguf) | Q5_0 | 0.71GB |
|
33 |
+
| [TinyLlama-1.1B-Tele.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q5_K_S.gguf) | Q5_K_S | 0.71GB |
|
34 |
+
| [TinyLlama-1.1B-Tele.Q5_K.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q5_K.gguf) | Q5_K | 0.73GB |
|
35 |
+
| [TinyLlama-1.1B-Tele.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q5_K_M.gguf) | Q5_K_M | 0.73GB |
|
36 |
+
| [TinyLlama-1.1B-Tele.Q5_1.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q5_1.gguf) | Q5_1 | 0.77GB |
|
37 |
+
| [TinyLlama-1.1B-Tele.Q6_K.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q6_K.gguf) | Q6_K | 0.84GB |
|
38 |
+
| [TinyLlama-1.1B-Tele.Q8_0.gguf](https://huggingface.co/RichardErkhov/AliMaatouk_-_TinyLlama-1.1B-Tele-gguf/blob/main/TinyLlama-1.1B-Tele.Q8_0.gguf) | Q8_0 | 1.09GB |
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
|
43 |
+
Original model description:
|
44 |
+
---
|
45 |
+
license: apache-2.0
|
46 |
+
language:
|
47 |
+
- en
|
48 |
+
pipeline_tag: text-generation
|
49 |
+
tags:
|
50 |
+
- nlp
|
51 |
+
---
|
52 |
+
|
53 |
+
# TinyLlama-1.1B-Tele Model Card
|
54 |
+
|
55 |
+
## Model Summary
|
56 |
+
|
57 |
+
The language model TinyLlama-1.1B-Tele is a Transformer with **1.1 billion** parameters, specialized in telecommunications. It is based on [TinyLlama-1.1B](https://huggingface.co/TinyLlama/TinyLlama_v1.1) and was continutally pretrained on [Tele-Data](https://huggingface.co/datasets/AliMaatouk/Tele-Data), a large-scale dataset of approximately 2.5 billion tokens of telecommunications material, including articles, standards, and general web content related to the telecommunications domain.
|
58 |
+
|
59 |
+
When assessed against telecommunications benchmarks such as [Tele-Eval](https://huggingface.co/datasets/AliMaatouk/Tele-Eval), TinyLlama-1.1B-Tele outperforms [TinyLlama-1.1B](https://huggingface.co/TinyLlama/TinyLlama_v1.1) by several percentage points. Additionally, TinyLlama-1.1B-Tele matches [TinyLlama-1.1B](https://huggingface.co/TinyLlama/TinyLlama_v1.1) across benchmarks related to common sense, language understanding, and logical reasoning. Thus, this adaptation was achieved with minimal compromise in performance on the original version.
|
60 |
+
|
61 |
+
### Context Length
|
62 |
+
|
63 |
+
The model was trained on a context length of 2048 tokens.
|
64 |
+
|
65 |
+
## Usage
|
66 |
+
|
67 |
+
TinyLlama-1.1B-Tele is a base model best suited for fine-tuning on applications related to telecommunications. It has not been fine-tuned to follow instructions and operates solely within a text completion framework. An example of this completion can be found below:
|
68 |
+
|
69 |
+
```markdown
|
70 |
+
Prompt: Shannon capacity is
|
71 |
+
|
72 |
+
Model: the capacity of a noiseless communication channel with a memoryless source. The Shannon capacity is a measure of the information rate that can be reliably transmitted over a noiseless channel.
|
73 |
+
```
|
74 |
+
|
75 |
+
The instruct version of this model can be found by following the link [TinyLlama-1.1B-Tele-it](https://huggingface.co/AliMaatouk/TinyLlama-1.1B-Tele-it).
|
76 |
+
|
77 |
+
## Sample Code
|
78 |
+
|
79 |
+
Below we share some code snippets on how to get quickly started with running the model. First, make sure to `pip install transformers`, then copy the snippet corresponding to your hardware and adapt it to your usecase.
|
80 |
+
|
81 |
+
#### Running the model on a CPU
|
82 |
+
|
83 |
+
|
84 |
+
```python
|
85 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
86 |
+
|
87 |
+
model = AutoModelForCausalLM.from_pretrained("AliMaatouk/TinyLlama-1.1B-Tele", torch_dtype="auto")
|
88 |
+
tokenizer = AutoTokenizer.from_pretrained("AliMaatouk/TinyLlama-1.1B-Tele")
|
89 |
+
|
90 |
+
prompt = "Shannon capacity is"
|
91 |
+
input_ids = tokenizer(prompt, return_tensors="pt")
|
92 |
+
outputs = model.generate(**input_ids, max_new_tokens=100)
|
93 |
+
|
94 |
+
generated_tokens = outputs[0, len(input_ids['input_ids'][0]):]
|
95 |
+
response = tokenizer.decode(generated_tokens, skip_special_tokens=True)
|
96 |
+
print(response)
|
97 |
+
```
|
98 |
+
|
99 |
+
#### Running the model on a single / multi GPU
|
100 |
+
|
101 |
+
```python
|
102 |
+
import torch
|
103 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
104 |
+
|
105 |
+
model = AutoModelForCausalLM.from_pretrained("AliMaatouk/TinyLlama-1.1B-Tele", torch_dtype="auto", device_map="auto")
|
106 |
+
tokenizer = AutoTokenizer.from_pretrained("AliMaatouk/TinyLlama-1.1B-Tele")
|
107 |
+
|
108 |
+
prompt = "Shannon capacity is"
|
109 |
+
input_ids = tokenizer(prompt, return_tensors="pt").to("cuda")
|
110 |
+
outputs = model.generate(**input_ids, max_new_tokens=100)
|
111 |
+
|
112 |
+
generated_tokens = outputs[0, len(input_ids['input_ids'][0]):]
|
113 |
+
response = tokenizer.decode(generated_tokens, skip_special_tokens=True)
|
114 |
+
print(response)
|
115 |
+
```
|
116 |
+
|
117 |
+
## Citation
|
118 |
+
|
119 |
+
You can find the paper with all details about the model at https://arxiv.org/abs/2409.05314. Please cite it as follows:
|
120 |
+
|
121 |
+
```bib
|
122 |
+
@misc{maatouk2024telellmsseriesspecializedlarge,
|
123 |
+
title={Tele-LLMs: A Series of Specialized Large Language Models for Telecommunications},
|
124 |
+
author={Ali Maatouk and Kenny Chirino Ampudia and Rex Ying and Leandros Tassiulas},
|
125 |
+
year={2024},
|
126 |
+
eprint={2409.05314},
|
127 |
+
archivePrefix={arXiv},
|
128 |
+
primaryClass={cs.IT},
|
129 |
+
url={https://arxiv.org/abs/2409.05314},
|
130 |
+
}
|
131 |
+
```
|
132 |
+
|