vrr/ProxyAI
2
0
Fork 0
mirror of https://github.com/carlrobertoh/ProxyAI.git synced 2026-05-10 12:10:14 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Add deepseek coder instruct models (1-33B)

Browse source
This commit is contained in:
Carl-Robert Linnupuu 2023-11-23 17:22:48 +02:00
parent 0faf48ed7e
commit 3797126de4
2 changed files with 27 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

10
src/main/java/ee/carlrobert/codegpt/completions/HuggingFaceModel.java
View file

@ -21,6 +21,16 @@ public enum HuggingFaceModel {
CODE_BOOGA_34B_Q4(34, 4, "CodeBooga-34B-v0.1-GGUF"),
CODE_BOOGA_34B_Q5(34, 5, "CodeBooga-34B-v0.1-GGUF"),
DEEPSEEK_CODER_1_3B_Q3(1, 3, "deepseek-coder-1.3b-instruct-GGUF"),
DEEPSEEK_CODER_1_3B_Q4(1, 4, "deepseek-coder-1.3b-instruct-GGUF"),
DEEPSEEK_CODER_1_3B_Q5(1, 5, "deepseek-coder-1.3b-instruct-GGUF"),
DEEPSEEK_CODER_6_7B_Q3(7, 3, "deepseek-coder-6.7b-instruct-GGUF"),
DEEPSEEK_CODER_6_7B_Q4(7, 4, "deepseek-coder-6.7b-instruct-GGUF"),
DEEPSEEK_CODER_6_7B_Q5(7, 5, "deepseek-coder-6.7b-instruct-GGUF"),
DEEPSEEK_CODER_33B_Q3(33, 3, "deepseek-coder-33b-instruct-GGUF"),
DEEPSEEK_CODER_33B_Q4(33, 4, "deepseek-coder-33b-instruct-GGUF"),
DEEPSEEK_CODER_33B_Q5(33, 5, "deepseek-coder-33b-instruct-GGUF"),
PHIND_CODE_LLAMA_34B_Q3(34, 3, "Phind-CodeLlama-34B-v2-GGUF"),
PHIND_CODE_LLAMA_34B_Q4(34, 4, "Phind-CodeLlama-34B-v2-GGUF"),
PHIND_CODE_LLAMA_34B_Q5(34, 5, "Phind-CodeLlama-34B-v2-GGUF"),

17
src/main/java/ee/carlrobert/codegpt/completions/llama/LlamaModel.java
View file

@ -38,6 +38,23 @@ public enum LlamaModel {
HuggingFaceModel.CODE_BOOGA_34B_Q3,
HuggingFaceModel.CODE_BOOGA_34B_Q4,
HuggingFaceModel.CODE_BOOGA_34B_Q5)),
DEEPSEEK_CODER(
"Deepseek Coder",
"Deepseek Coder is composed of a series of code language models, each trained "
+ "from scratch on 2T tokens, with a composition of 87% code and 13% natural language "
+ "in both English and Chinese. It achieves state-of-the-art performance among "
+ "open-source code models on multiple programming languages and various benchmarks.",
PromptTemplate.ALPACA,
List.of(
HuggingFaceModel.DEEPSEEK_CODER_1_3B_Q3,
HuggingFaceModel.DEEPSEEK_CODER_1_3B_Q4,
HuggingFaceModel.DEEPSEEK_CODER_1_3B_Q5,
HuggingFaceModel.DEEPSEEK_CODER_6_7B_Q3,
HuggingFaceModel.DEEPSEEK_CODER_6_7B_Q4,
HuggingFaceModel.DEEPSEEK_CODER_6_7B_Q5,
HuggingFaceModel.DEEPSEEK_CODER_33B_Q3,
HuggingFaceModel.DEEPSEEK_CODER_33B_Q4,
HuggingFaceModel.DEEPSEEK_CODER_33B_Q5)),
PHIND_CODE_LLAMA(
"Phind Code Llama",
"This model is fine-tuned from Phind-CodeLlama-34B-v1 on an additional 1.5B tokens "