Add new ranked GGMLv3 models (#417)

* Feature: add some best ranked models * Fix: models size * Update README.md --------- Co-authored-by: pabl-o-ce <[email protected]> Co-authored-by: Juan Calderon-Perez <[email protected]>
serge-chat · Jun 12, 2023 · 6bf84dd · 6bf84dd
1 parent 4970865
commit 6bf84dd
Show file tree

Hide file tree

Showing 2 changed files with 171 additions and 86 deletions.
diff --git a/README.md b/README.md
@@ -1,4 +1,4 @@
-# 🦙 Serge - LLaMA made easy
+# Serge - LLaMA made easy 🦙
 
 ![License](https://img.shields.io/github/license/serge-chat/serge)
 [![Discord](https://img.shields.io/discord/1088427963801948201?label=Discord)](https://discord.gg/62Hc6FEYQH)
@@ -49,17 +49,24 @@ Then, just visit http://localhost:8008/, You can find the API documentation at h
 
 Ensure you have Docker Desktop installed, WSL2 configured, and enough free RAM to run models. 
 
-## ☁️Kubernetes & Docker Compose Setup
+## ☁️ Kubernetes & Docker Compose Setup
 
 Instructions for setting up Serge on Kubernetes can be found in the [wiki](https://github.com/serge-chat/serge/wiki/Integrating-Serge-in-your-orchestration#kubernetes-example).
 
 ## 🧠 Supported Models
 
 We currently support the following models:
 
+- Airoboros 🎈
+  - Airoboros-7B
+  - Airoboros-13B
+  - Airoboros-30B
 - Alpaca 🦙
   - Alpaca-LoRA-65B
   - GPT4-Alpaca-LoRA-30B
+- Chronos 🌑
+  - Chronos-13B
+  - Chronos-33B
 - GPT4All 🌍
   - GPT4All-13B
 - Guanaco 🦙
@@ -70,6 +77,9 @@ We currently support the following models:
 - Koala 🐨
   - Koala-7B
   - Koala-13B
+- Llama 🦙
+  - FinLlama-33B
+  - Llama-Supercot-30B
 - Lazarus 💀
   - Lazarus-30B
 - Nous 🧠
@@ -80,13 +90,18 @@ We currently support the following models:
   - Samantha-7B
   - Samantha-13B
   - Samantha-33B
-- Stable 🐎
-  - Stable-Vicuna-13B
+- Tulu 🎚
+  - Tulu-7B
+  - Tulu-13B
+  - Tulu-30B
 - Vicuna 🦙
+  - Stable-Vicuna-13B
   - Vicuna-CoT-7B
   - Vicuna-CoT-13B
   - Vicuna-v1.1-7B
   - Vicuna-v1.1-13B
+  - VicUnlocked-30B
+  - VicUnlocked-65B
 - Wizard 🧙
   - Wizard-Mega-13B
   - Wizard-Vicuna-Uncensored-7B

diff --git a/api/src/serge/routers/model.py b/api/src/serge/routers/model.py
@@ -13,6 +13,26 @@
 )
 
 models_info = {
+    "Airoboros-7B": [
+        "TheBloke/airoboros-7b-gpt4-GGML",
+        "airoboros-7b-gpt4.ggmlv3.q5_1.bin",
+        5.06e9,
+    ],
+    "Airoboros-7B-q6_K": [
+        "TheBloke/airoboros-7b-gpt4-GGML",
+        "airoboros-7b-gpt4.ggmlv3.q6_K.bin",
+        5.53e9,
+    ],
+    "Airoboros-13B": [
+        "TheBloke/airoboros-13b-gpt4-GGML",
+        "airoboros-13b-gpt4.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "Airoboros-33B": [
+        "TheBloke/airoboros-33b-gpt4-GGML",
+        "airoboros-33b-gpt4.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
     "GPT4AlpacaLoRA-30B": [
         "TheBloke/gpt4-alpaca-lora-30B-4bit-GGML",
         "gpt4-alpaca-lora-30b.ggmlv3.q5_1.bin",
@@ -23,34 +43,129 @@
         "alpaca-lora-65B.ggmlv3.q5_K_M.bin",
         46.2e9,
     ],
-    "OpenAssistant-30B": [
-        "TheBloke/OpenAssistant-SFT-7-Llama-30B-GGML",
-        "OpenAssistant-SFT-7-Llama-30B.ggmlv3.q6_K.bin",
-        26.7e9,
-    ],
     "GPT4All-13B": [
         "TheBloke/GPT4All-13B-snoozy-GGML",
         "GPT4All-13B-snoozy.ggmlv3.q6_K.bin",
         10.7e9,
     ],
-    "StableVicuna-13B": [
-        "TheBloke/stable-vicuna-13B-GGML",
-        "stable-vicuna-13B.ggmlv3.q6_K.bin",
+    "Chronos-13B": [
+        "TheBloke/chronos-13B-GGML",
+        "chronos-13b.ggmlv3.q6_K.bin",
         10.7e9,
     ],
-    "Vicuna-v1.1-7B": [
-        "TheBloke/vicuna-7B-1.1-GGML",
-        "vicuna-7b-1.1.ggmlv3.q5_1.bin",
+    "Chronos-33B": [
+        "TheBloke/chronos-33b-GGML",
+        "chronos-33b.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Guanaco-7B": [
+        "TheBloke/guanaco-7B-GGML",
+        "guanaco-7B.ggmlv3.q5_1.bin",
         5.06e9,
     ],
-    "Vicuna-v1.1-7B-q6_K": [
-        "TheBloke/vicuna-7B-1.1-GGML",
-        "vicuna-7b-1.1.ggmlv3.q6_K.bin",
+    "Guanaco-7B-q6_K": [
+        "TheBloke/guanaco-7B-GGML",
+        "guanaco-7B.ggmlv3.q6_K.bin",
         5.53e9,
     ],
-    "Vicuna-v1.1-13B": [
-        "TheBloke/vicuna-13b-1.1-GGML",
-        "vicuna-13b-1.1.ggmlv3.q6_K.bin",
+    "Guanaco-13B": [
+        "TheBloke/guanaco-13B-GGML",
+        "guanaco-13B.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "Guanaco-33B": [
+        "TheBloke/guanaco-33B-GGML",
+        "guanaco-33B.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Guanaco-65B": [
+        "TheBloke/guanaco-65B-GGML",
+        "guanaco-65B.ggmlv3.q5_K_M.bin",
+        46.2e9,
+    ],
+    "Koala-7B": [
+        "TheBloke/koala-7B-GGML",
+        "koala-7B.ggmlv3.q5_1.bin",
+        5.06e9,
+    ],
+    "Koala-7B-q6_K": [
+        "TheBloke/koala-7B-GGML",
+        "koala-7B.ggmlv3.q6_K.bin",
+        5.53e9,
+    ],
+    "Koala-13B": [
+        "TheBloke/koala-13B-GGML",
+        "koala-13B.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "FinLlama-33B": [
+        "TheBloke/fin-llama-33B-GGML",
+        "fin-llama-33b.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Llama-Supercot-30B": [
+        "TheBloke/llama-30b-supercot-GGML",
+        "llama-30b-supercot.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Lazarus-30B": [
+        "TheBloke/30B-Lazarus-GGML",
+        "30b-Lazarus.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Nous-Hermes-13B": [
+        "TheBloke/Nous-Hermes-13B-GGML",
+        "nous-hermes-13b.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "OpenAssistant-30B": [
+        "TheBloke/OpenAssistant-SFT-7-Llama-30B-GGML",
+        "OpenAssistant-SFT-7-Llama-30B.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Samantha-7B": [
+        "TheBloke/Samantha-7B-GGML",
+        "Samantha-7B.ggmlv3.q5_1.bin",
+        5.06e9,
+    ],
+    "Samantha-7B-q6_K": [
+        "TheBloke/Samantha-7B-GGML",
+        "Samantha-7B.ggmlv3.q6_K.bin",
+        5.53e9,
+    ],
+    "Samantha-13B": [
+        "TheBloke/Samantha-13B-GGML",
+        "samantha-13b.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "Samantha-33B": [
+        "TheBloke/Samantha-33B-GGML",
+        "samantha-33B.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "Tulu-7B": [
+        "TheBloke/tulu-7B-GGML",
+        "tulu-7b.ggmlv3.q5_1.bin",
+        5.06e9,
+    ],
+    "Tulu-7B-q6_K": [
+        "TheBloke/tulu-7B-GGML",
+        "tulu-7b.ggmlv3.q6_K.bin",
+        5.06e9,
+    ],
+    "Tulu-13B": [
+        "TheBloke/tulu-13B-GGML",
+        "tulu-13b.ggmlv3.q6_K.bin",
+        10.7e9,
+    ],
+    "Tulu-30B": [
+        "TheBloke/tulu-30B-GGML",
+        "tulu-30b.ggmlv3.q6_K.bin",
+        26.7e9,
+    ],
+    "StableVicuna-13B": [
+        "TheBloke/stable-vicuna-13B-GGML",
+        "stable-vicuna-13B.ggmlv3.q6_K.bin",
         10.7e9,
     ],
     "Vicuna-CoT-7B": [
@@ -68,31 +183,36 @@
         "vicuna-13b-cot.ggmlv3.q6_K.bin",
         10.7e9,
     ],
-    "Guanaco-7B": [
-        "TheBloke/guanaco-7B-GGML",
-        "guanaco-7B.ggmlv3.q5_1.bin",
+    "Vicuna-v1.1-7B": [
+        "TheBloke/vicuna-7B-1.1-GGML",
+        "vicuna-7b-1.1.ggmlv3.q5_1.bin",
         5.06e9,
     ],
-    "Guanaco-7B-q6_K": [
-        "TheBloke/guanaco-7B-GGML",
-        "guanaco-7B.ggmlv3.q6_K.bin",
+    "Vicuna-v1.1-7B-q6_K": [
+        "TheBloke/vicuna-7B-1.1-GGML",
+        "vicuna-7b-1.1.ggmlv3.q6_K.bin",
         5.53e9,
     ],
-    "Guanaco-13B": [
-        "TheBloke/guanaco-13B-GGML",
-        "guanaco-13B.ggmlv3.q6_K.bin",
+    "Vicuna-v1.1-13B": [
+        "TheBloke/vicuna-13b-1.1-GGML",
+        "vicuna-13b-1.1.ggmlv3.q6_K.bin",
         10.7e9,
     ],
-    "Guanaco-33B": [
-        "TheBloke/guanaco-33B-GGML",
-        "guanaco-33B.ggmlv3.q6_K.bin",
+    "VicUnlocked-30B": [
+        "TheBloke/VicUnlocked-30B-LoRA-GGML",
+        "VicUnlocked-30B-LoRA.ggmlv3.q6_K.bin",
         26.7e9,
     ],
-    "Guanaco-65B": [
-        "TheBloke/guanaco-65B-GGML",
-        "guanaco-65B.ggmlv3.q5_K_M.bin",
+    "VicUnlocked-65B": [
+        "TheBloke/VicUnlocked-alpaca-65B-QLoRA-GGML",
+        "vicunlocked-65b.ggmlv3.q5_K_M.bin",
         46.2e9,
     ],
+    "Wizard-Mega-13B": [
+        "TheBloke/wizard-mega-13B-GGML",
+        "wizard-mega-13B.ggmlv3.q5_1.bin",
+        9.76e9,
+    ],
     "Wizard-Vicuna-Uncensored-7B": [
         "TheBloke/Wizard-Vicuna-7B-Uncensored-GGML",
         "Wizard-Vicuna-7B-Uncensored.ggmlv3.q5_1.bin",
@@ -138,56 +258,6 @@
         "WizardLM-30B-Uncensored.ggmlv3.q6_K.bin",
         26.7e9,
     ],
-    "Wizard-Mega-13B": [
-        "TheBloke/wizard-mega-13B-GGML",
-        "wizard-mega-13B.ggmlv3.q5_1.bin",
-        9.76e9,
-    ],
-    "Lazarus-30B": [
-        "TheBloke/30B-Lazarus-GGML",
-        "30b-Lazarus.ggmlv3.q6_K.bin",
-        26.7e9,
-    ],
-    "Nous-Hermes-13B": [
-        "TheBloke/Nous-Hermes-13B-GGML",
-        "nous-hermes-13b.ggmlv3.q6_K.bin",
-        10.7e9,
-    ],
-    "Samantha-7B": [
-        "TheBloke/Samantha-7B-GGML",
-        "Samantha-7B.ggmlv3.q5_1.bin",
-        5.06e9,
-    ],
-    "Samantha-7B-q6_K": [
-        "TheBloke/Samantha-7B-GGML",
-        "Samantha-7B.ggmlv3.q6_K.bin",
-        5.53e9,
-    ],
-    "Samantha-13B": [
-        "TheBloke/Samantha-13B-GGML",
-        "samantha-13b.ggmlv3.q6_K.bin",
-        10.7e9,
-    ],
-    "Samantha-33B": [
-        "TheBloke/Samantha-33B-GGML",
-        "samantha-33B.ggmlv3.q6_K.bin",
-        26.7e9,
-    ],
-    "Koala-7B": [
-        "TheBloke/koala-7B-GGML",
-        "koala-7B.ggmlv3.q5_1.bin",
-        5.06e9,
-    ],
-    "Koala-7B-q6_K": [
-        "TheBloke/koala-7B-GGML",
-        "koala-7B.ggmlv3.q6_K.bin",
-        5.53e9,
-    ],
-    "Koala-13B": [
-        "TheBloke/koala-13B-GGML",
-        "koala-13B.ggmlv3.q6_K.bin",
-        10.7e9,
-    ],
 }
 
 WEIGHTS = "/usr/src/app/weights/"