Update README.md
Browse files
README.md
CHANGED
|
@@ -12,8 +12,6 @@ datasets:
|
|
| 12 |
|
| 13 |
### This repository contains quantized conversions of EleutherAI's Pythia Deduped checkpoints.
|
| 14 |
|
| 15 |
-
If you're starting off, I highly recommend for you to get models from the newest directory [(2023-05-25)](https://huggingface.co/Merry/ggml-pythia-deduped/tree/main/2023-05-25).
|
| 16 |
-
|
| 17 |
[Click here if you're looking for ggmlv1 and ggmlv2 models.](https://huggingface.co/Merry/ggml-pythia-deduped/tree/a695a4c30c01ed9a41200c01f85d47c819fc93dd).
|
| 18 |
|
| 19 |
# RAM USAGE
|
|
@@ -36,19 +34,6 @@ ggmlv3-pythia-1.4b-deduped-q5_1.bin | 1.3 GiB
|
|
| 36 |
ggmlv3-pythia-2.8b-deduped-q5_1.bin | 2.3 GiB
|
| 37 |
|
| 38 |
*Tested on KoboldCpp with OpenBLAS enabled.*
|
| 39 |
-
|
| 40 |
-
**Versions:**
|
| 41 |
-
|
| 42 |
-
**2023-04-20:** *q4_3. Used [commit 05f3079](https://github.com/ggerganov/ggml/tree/05f307971862b83df12fada0c42ee027ba5a82b5/examples/stablelm)*
|
| 43 |
-
|
| 44 |
-
**2023-04-30:** *q5_0, q5_1, and q8_0, up to 2.8B. I can't upload all conversions of 6.9B and 12B due to my internet. Used [commit 5dd92f4](https://github.com/ggerganov/ggml/tree/5dd92f421ee44f18b8fde0afbf5ca8fc7bf93841/examples/stablelm)*
|
| 45 |
-
|
| 46 |
-
**2023-05-06:** *q4_0 and q4_2, up to 2.8B. Used [commit ff6e03c](https://github.com/ggerganov/ggml/tree/ff6e03cbcd9bf6e9fa41d49f2495c042efae4dc6/examples/stablelm)*
|
| 47 |
-
|
| 48 |
-
**2023-05-15:** *New quantization format (ggmlv2). q4_0 and q5_1, up to 2.8B. Used [commit 010203f](https://github.com/ggerganov/ggml/tree/010203f94a85df5c86b773dc5acb698c8e7b1e7b/examples/gpt-neox)*
|
| 49 |
-
|
| 50 |
-
**2023-05-25:** *New quantization format (ggmlv3). q4_0 and q5_1, up to 2.8B. Used [commit 73ad593](https://github.com/ggerganov/ggml/tree/73ad593cf84f864f0fcfd3a196253575c70d66a2/examples/gpt-neox)*
|
| 51 |
-
|
| 52 |
**Notes:**
|
| 53 |
- The models have been converted with ggerganov/ggml's gpt-neox conversion script, and tested only on KoboldCpp. Other frontends that support GGML-based conversions of GPT-NeoX *should* work, but I can't promise anything.
|
| 54 |
- They're sorted by date based on when they were converted so it was easier to track breaking changes. If you're just starting off I highly recommend the latest, which is currently 2023-05-25. Combined with KoboldCpp v1.25.1+ this improved the tokenizer, which in my testing reduces occurrences of broken words like "Alicae" or "Reimu Hai-ku-rei".
|
|
|
|
| 12 |
|
| 13 |
### This repository contains quantized conversions of EleutherAI's Pythia Deduped checkpoints.
|
| 14 |
|
|
|
|
|
|
|
| 15 |
[Click here if you're looking for ggmlv1 and ggmlv2 models.](https://huggingface.co/Merry/ggml-pythia-deduped/tree/a695a4c30c01ed9a41200c01f85d47c819fc93dd).
|
| 16 |
|
| 17 |
# RAM USAGE
|
|
|
|
| 34 |
ggmlv3-pythia-2.8b-deduped-q5_1.bin | 2.3 GiB
|
| 35 |
|
| 36 |
*Tested on KoboldCpp with OpenBLAS enabled.*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
**Notes:**
|
| 38 |
- The models have been converted with ggerganov/ggml's gpt-neox conversion script, and tested only on KoboldCpp. Other frontends that support GGML-based conversions of GPT-NeoX *should* work, but I can't promise anything.
|
| 39 |
- They're sorted by date based on when they were converted so it was easier to track breaking changes. If you're just starting off I highly recommend the latest, which is currently 2023-05-25. Combined with KoboldCpp v1.25.1+ this improved the tokenizer, which in my testing reduces occurrences of broken words like "Alicae" or "Reimu Hai-ku-rei".
|