Update README.md
Browse files
README.md
CHANGED
|
@@ -25,6 +25,61 @@ Each branch contains an individual bits per weight, with the main one containing
|
|
| 25 |
|
| 26 |
<a href="https://huggingface.co/machinez/zephyr-orpo-141b-A35b-v0.1-exl2/tree/2_75">2.75 bits per weight - Fits Quad Nvidia Tesla P100 16gb at 16k context</a>
|
| 27 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
## Download instructions
|
| 29 |
|
| 30 |
With git:
|
|
|
|
| 25 |
|
| 26 |
<a href="https://huggingface.co/machinez/zephyr-orpo-141b-A35b-v0.1-exl2/tree/2_75">2.75 bits per weight - Fits Quad Nvidia Tesla P100 16gb at 16k context</a>
|
| 27 |
|
| 28 |
+
## Sample instructions to load in TabbyAPI @ 1.5bpw on 3x Nvidia Tesla P100 16gb at 4k context
|
| 29 |
+
```JSON
|
| 30 |
+
{
|
| 31 |
+
"name": "Machinez_zephyr-orpo-141b-A35b-v0.1_1.5bpw",
|
| 32 |
+
"max_seq_len": 4096,
|
| 33 |
+
"override_base_seq_len": 4096,
|
| 34 |
+
"gpu_split_auto": false,
|
| 35 |
+
"autosplit_reserve": [
|
| 36 |
+
96
|
| 37 |
+
],
|
| 38 |
+
"gpu_split": [
|
| 39 |
+
14.15,
|
| 40 |
+
14,
|
| 41 |
+
15
|
| 42 |
+
],
|
| 43 |
+
"rope_scale": 1,
|
| 44 |
+
"rope_alpha": 1,
|
| 45 |
+
"no_flash_attention": false,
|
| 46 |
+
"cache_mode": "fp16",
|
| 47 |
+
"prompt_template": "string",
|
| 48 |
+
"num_experts_per_token": 0,
|
| 49 |
+
"use_cfg": true,
|
| 50 |
+
"fasttensors": false,
|
| 51 |
+
"skip_queue": false
|
| 52 |
+
}
|
| 53 |
+
```
|
| 54 |
+
|
| 55 |
+
## Sample instructions to load in TabbyAPI @ 2.75bpw on 4x Nvidia Tesla P100 16gb at 16k context
|
| 56 |
+
```JSON
|
| 57 |
+
{
|
| 58 |
+
"name": "Machinez_zephyr-orpo-141b-A35b-v0.1_2.75bpw",
|
| 59 |
+
"max_seq_len": 16384,
|
| 60 |
+
"override_base_seq_len": 16384,
|
| 61 |
+
"gpu_split_auto": false,
|
| 62 |
+
"autosplit_reserve": [
|
| 63 |
+
96
|
| 64 |
+
],
|
| 65 |
+
"gpu_split": [
|
| 66 |
+
12.5,
|
| 67 |
+
13,
|
| 68 |
+
13,
|
| 69 |
+
16.1
|
| 70 |
+
],
|
| 71 |
+
"rope_scale": 1,
|
| 72 |
+
"rope_alpha": 1,
|
| 73 |
+
"no_flash_attention": false,
|
| 74 |
+
"cache_mode": "fp16",
|
| 75 |
+
"prompt_template": "string",
|
| 76 |
+
"num_experts_per_token": 0,
|
| 77 |
+
"use_cfg": true,
|
| 78 |
+
"fasttensors": false,
|
| 79 |
+
"skip_queue": false
|
| 80 |
+
}
|
| 81 |
+
```
|
| 82 |
+
|
| 83 |
## Download instructions
|
| 84 |
|
| 85 |
With git:
|