Text Generation
Transformers
code
Eval Results
bartowski commited on
Commit
ad21e31
·
verified ·
1 Parent(s): e36cb0a

Update VRAM table

Browse files
Files changed (1) hide show
  1. README.md +7 -13
README.md CHANGED
@@ -73,22 +73,16 @@ Using <a href="https://github.com/turboderp/exllamav2/">turboderp's ExLlamaV2 v0
73
 
74
  Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
75
 
76
- Conversion was done using the default calibration dataset.
77
-
78
- Default arguments used except when the bits per weight is above 6.0, at that point the lm_head layer is quantized at 8 bits per weight instead of the default 6.
79
-
80
  Original model: https://huggingface.co/bigcode/starcoder2-15b
81
 
82
 
83
- <a href="https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/8_0">8.0 bits per weight</a>
84
-
85
- <a href="https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/6_5">6.5 bits per weight</a>
86
-
87
- <a href="https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/5_0">5.0 bits per weight</a>
88
-
89
- <a href="https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/4_25">4.25 bits per weight</a>
90
-
91
- <a href="https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/3_5">3.5 bits per weight</a>
92
 
93
 
94
  ## Download instructions
 
73
 
74
  Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
75
 
 
 
 
 
76
  Original model: https://huggingface.co/bigcode/starcoder2-15b
77
 
78
 
79
+ | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
80
+ | ----- | ---- | ------- | ------ | ------ | ------ | ------------ |
81
+ | [8_0](https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/8_0) | 8.0 | 8.0 | 16.6 GB | 17.5 GB | 18.8 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
82
+ | [6_5](https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/6_5) | 6.5 | 8.0 | 13.9 GB | 14.9 GB | 16.2 GB | Near unquantized performance at vastly reduced size, **recommended**. |
83
+ | [5_0](https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/5_0) | 5.0 | 6.0 | 11.2 GB | 12.2 GB | 13.5 GB | Slightly lower quality vs 6.5. |
84
+ | [4_25](https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/4_25) | 4.25 | 6.0 | 9.8 GB | 10.7 GB | 12.0 GB | GPTQ equivalent bits per weight. |
85
+ | [3_5](https://huggingface.co/bartowski/starcoder2-15b-exl2/tree/3_5) | 3.5 | 6.0 | 8.4 GB | 9.3 GB | 10.6 GB | Lower quality, not recommended. |
 
 
86
 
87
 
88
  ## Download instructions