doberst commited on
Commit
44425c9
1 Parent(s): da3f4f9

Upload 2 files

Browse files
Files changed (2) hide show
  1. README.md +36 -2
  2. config.json +28 -0
README.md CHANGED
@@ -1,4 +1,38 @@
1
  ---
2
  license: gemma
3
- inference: false
4
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: gemma
3
+ inference: false
4
+ tags:
5
+ - green
6
+ - p2
7
+ - llmware-chat
8
+ - ov
9
+ ---
10
+
11
+ # gemma-2b-it-ov
12
+
13
+ **gemma-2b-it-ov-ov** is an OpenVino int4 quantized version of Google's Gemma-2B with Instruct Training (IT), providing a very fast, very small inference implementation, optimized for AI PCs using Intel GPU, CPU and NPU.
14
+
15
+ [**gemma-2b-it-ov**](https://huggingface.co/google/gemma-2b-it) is a leading open source foundation model from Google.
16
+
17
+
18
+ ### Model Description
19
+
20
+ - **Developed by:** Google
21
+ - **Quantized by:** llmware
22
+ - **Model type:** gemma-2b
23
+ - **Parameters:** 2 billion
24
+ - **Model Parent:** google/gemma-2b-it
25
+ - **Language(s) (NLP):** English
26
+ - **License:** Apache 2.0
27
+ - **Uses:** General purpose chat
28
+ - **RAG Benchmark Accuracy Score:** NA
29
+ - **Quantization:** int4
30
+
31
+
32
+ ## Model Card Contact
33
+
34
+ [llmware on github](https://www.github.com/llmware-ai/llmware)
35
+
36
+ [llmware on hf](https://www.huggingface.co/llmware)
37
+
38
+ [llmware website](https://www.llmware.ai)
config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/gemma-2b-it",
3
+ "architectures": [
4
+ "GemmaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 2,
9
+ "eos_token_id": 1,
10
+ "head_dim": 256,
11
+ "hidden_act": "gelu",
12
+ "hidden_activation": "gelu_pytorch_tanh",
13
+ "hidden_size": 2048,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 16384,
16
+ "max_position_embeddings": 8192,
17
+ "model_type": "gemma",
18
+ "num_attention_heads": 8,
19
+ "num_hidden_layers": 18,
20
+ "num_key_value_heads": 1,
21
+ "pad_token_id": 0,
22
+ "rms_norm_eps": 1e-06,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "transformers_version": "4.43.4",
26
+ "use_cache": true,
27
+ "vocab_size": 256000
28
+ }