Downloads · 30 days
361
6% of all-time downloads
Defetya_-_gemma-2b-ru-gguf
RichardErkhov/Defetya_-_gemma-2b-ru-gguf
Defetya_-_gemma-2b-ru-gguf is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
All-time downloads
5.9K
Public
Repo size
30.7 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.gguf35.6 GB · 100%
At a glance
- Access
- Public
- Created
- Jul 22, 2024
- Updated
- Jul 22, 2024
- SHA
- ff0044dd
Model card
From the Hugging Face model README
Quantization made by Richard Erkhov.
gemma-2b-ru - GGUF
- Model creator: https://huggingface.co/Defetya/
- Original model: https://huggingface.co/Defetya/gemma-2b-ru/
| Name | Quant method | Size |
|---|---|---|
| gemma-2b-ru.Q2_K.gguf | Q2_K | 1.08GB |
| gemma-2b-ru.IQ3_XS.gguf | IQ3_XS | 1.16GB |
| gemma-2b-ru.IQ3_S.gguf | IQ3_S | 1.2GB |
| gemma-2b-ru.Q3_K_S.gguf | Q3_K_S | 1.2GB |
| gemma-2b-ru.IQ3_M.gguf | IQ3_M | 1.22GB |
| gemma-2b-ru.Q3_K.gguf | Q3_K | 1.29GB |
| gemma-2b-ru.Q3_K_M.gguf | Q3_K_M | 1.29GB |
| gemma-2b-ru.Q3_K_L.gguf | Q3_K_L | 1.36GB |
| gemma-2b-ru.IQ4_XS.gguf | IQ4_XS | 1.4GB |
| gemma-2b-ru.Q4_0.gguf | Q4_0 | 1.44GB |
| gemma-2b-ru.IQ4_NL.gguf | IQ4_NL | 1.45GB |
| gemma-2b-ru.Q4_K_S.gguf | Q4_K_S | 1.45GB |
| gemma-2b-ru.Q4_K.gguf | Q4_K | 1.52GB |
| gemma-2b-ru.Q4_K_M.gguf | Q4_K_M | 1.52GB |
| gemma-2b-ru.Q4_1.gguf | Q4_1 | 1.56GB |
| gemma-2b-ru.Q5_0.gguf | Q5_0 | 1.68GB |
| gemma-2b-ru.Q5_K_S.gguf | Q5_K_S | 1.68GB |
| gemma-2b-ru.Q5_K.gguf | Q5_K | 1.71GB |
| gemma-2b-ru.Q5_K_M.gguf | Q5_K_M | 1.71GB |
| gemma-2b-ru.Q5_1.gguf | Q5_1 | 1.79GB |
| gemma-2b-ru.Q6_K.gguf | Q6_K | 1.92GB |
| gemma-2b-ru.Q8_0.gguf | Q8_0 | 2.49GB |
Original model description:
license: apache-2.0 datasets:
- oscar-corpus/OSCAR-2301
This model is a result of second stage pre-training of Google's Gemma 2B (https://huggingface.co/google/gemma-2b) for roughly 150B tokens on the combination of English + Russian subset of oscar and wiki datasets.
This is a raw pre-trained model, created with further fine-tuning in mind. Goal of this project is to further research cross-linguistic capabilities of open-source LLMs and to create a strong open-source foundational LLM that would be fluent in Russian language. More about it will be in the upcoming blog and/or research paper.
This model was pre-trained using EasyLM's fork as a framework (JAX) on Google's v4-32 TPU which was generously provided under the TRC program. The model reached ~ 1.5 in training loss, LR was roughly 5e-5.
I'm planning on releasing a chat model that would ungergo full-parameter SFT and DPO on Ilya Gusev's datasets.
- Created
- Jul 22, 2024
- Updated
- Jul 22, 2024