From 18844081638585cf5d4bb95fd264d5555d3211da Mon Sep 17 00:00:00 2001
From: Nestor Qin <imba.qxy@gmail.com>
Date: Mon, 4 Nov 2024 15:04:51 -0500
Subject: [PATCH] feat: update SmolLM to SmolLM2 models

---
 app/constant.ts     | 365 ++++++++++----------------------------------
 app/store/config.ts |   4 +-
 app/utils.ts        |  20 +++
 package.json        |   2 +-
 yarn.lock           |   8 +-
 5 files changed, 105 insertions(+), 294 deletions(-)

diff --git a/app/constant.ts b/app/constant.ts
index c37f2d17..1ec075a5 100644
--- a/app/constant.ts
+++ b/app/constant.ts
@@ -1,4 +1,6 @@
+import { prebuiltAppConfig } from "@mlc-ai/web-llm";
 import { ModelRecord } from "./client/api";
+import { getQuantization, getSize } from "./utils";
 
 export const OWNER = "mlc-ai";
 export const REPO = "web-llm-chat";
@@ -64,13 +66,12 @@ Latex inline format: \\(x^2\\)
 Latex block format: $$e=mc^2$$
 `;
 
-export const DEFAULT_MODELS: ModelRecord[] = [
+const DEFAULT_MODEL_BASES: ModelRecord[] = [
   // Phi-3.5 Vision
   {
     name: "Phi-3.5-vision-instruct-q4f32_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 3.5 Vision",
     recommended_config: {
       temperature: 1,
@@ -83,7 +84,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3.5-vision-instruct-q4f16_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 3.5 Vision",
     recommended_config: {
       temperature: 1,
@@ -97,8 +97,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-1B-Instruct-q4f32_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "1B",
-    quantization: "q4f32_1",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -111,8 +109,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-1B-Instruct-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "1B",
-    quantization: "q4f16_1",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -125,8 +121,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-1B-Instruct-q0f32-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "1B",
-    quantization: "q0f32",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -139,8 +133,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-1B-Instruct-q0f16-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "1B",
-    quantization: "q0f16",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -153,8 +145,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-3B-Instruct-q4f32_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "3B",
-    quantization: "q4f32_1",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -167,8 +157,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.2-3B-Instruct-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "3B",
-    quantization: "q4f16_1",
     family: "Llama 3.2",
     recommended_config: {
       temperature: 0.6,
@@ -182,8 +170,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.1-8B-Instruct-q4f32_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Llama 3.1",
     recommended_config: {
       temperature: 0.6,
@@ -196,8 +182,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.1-8B-Instruct-q4f16_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Llama 3.1",
     recommended_config: {
       temperature: 0.6,
@@ -210,8 +194,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.1-8B-Instruct-q4f32_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Llama 3.1",
     recommended_config: {
       temperature: 0.6,
@@ -224,8 +206,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.1-8B-Instruct-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Llama 3.1",
     recommended_config: {
       temperature: 0.6,
@@ -238,8 +218,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Hermes-2-Pro-Llama-3-8B-q4f16_1-MLC",
     display_name: "Hermes",
     provider: "NousResearch",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Hermes 2 Pro",
     recommended_config: {
       temperature: 1,
@@ -252,8 +230,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Hermes-2-Pro-Llama-3-8B-q4f32_1-MLC",
     display_name: "Hermes",
     provider: "NousResearch",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Hermes 2 Pro",
     recommended_config: {
       temperature: 1,
@@ -266,8 +242,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Hermes-3-Llama-3.1-8B-q4f32_1-MLC",
     display_name: "Hermes",
     provider: "NousResearch",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Hermes 3",
     recommended_config: {
       temperature: 0.6,
@@ -280,8 +254,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Hermes-3-Llama-3.1-8B-q4f16_1-MLC",
     display_name: "Hermes",
     provider: "NousResearch",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Hermes 3",
     recommended_config: {
       temperature: 0.6,
@@ -294,8 +266,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Hermes-2-Pro-Mistral-7B-q4f16_1-MLC",
     display_name: "Hermes",
     provider: "NousResearch",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Hermes 2 Pro",
     recommended_config: {
       temperature: 0.7,
@@ -308,7 +278,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3.5-mini-instruct-q4f16_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 3.5 Mini",
     recommended_config: {
       temperature: 1,
@@ -321,7 +290,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3.5-mini-instruct-q4f32_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 3.5 Mini",
     recommended_config: {
       temperature: 1,
@@ -334,7 +302,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3.5-mini-instruct-q4f16_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 3.5 Mini",
     recommended_config: {
       temperature: 1,
@@ -347,7 +314,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3.5-mini-instruct-q4f32_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 3.5 Mini",
     recommended_config: {
       temperature: 1,
@@ -360,8 +326,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Mistral-7B-Instruct-v0.3-q4f16_1-MLC",
     display_name: "Mistral",
     provider: "Mistral AI",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Mistral",
     recommended_config: {
       temperature: 1,
@@ -374,8 +338,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Mistral-7B-Instruct-v0.3-q4f32_1-MLC",
     display_name: "Mistral",
     provider: "Mistral AI",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Mistral",
     recommended_config: {
       temperature: 1,
@@ -388,8 +350,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Mistral-7B-Instruct-v0.2-q4f16_1-MLC",
     display_name: "Mistral",
     provider: "Mistral AI",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Mistral",
     recommended_config: {
       temperature: 0.7,
@@ -400,8 +360,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "OpenHermes-2.5-Mistral-7B-q4f16_1-MLC",
     display_name: "OpenHermes",
     provider: "NousResearch",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Hermes",
     recommended_config: {
       temperature: 0.7,
@@ -412,8 +370,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "NeuralHermes-2.5-Mistral-7B-q4f16_1-MLC",
     display_name: "NeuralHermes",
     provider: "Maxime Labonne",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Hermes",
     recommended_config: {
       temperature: 0.7,
@@ -424,164 +380,113 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "WizardMath-7B-V1.1-q4f16_1-MLC",
     display_name: "WizardMath",
     provider: "WizardLM",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "WizardMath",
     recommended_config: {
       temperature: 0.7,
       top_p: 0.95,
     },
   },
+  // SmolLM2
   {
-    name: "SmolLM-1.7B-Instruct-q0f16-MLC",
+    name: "SmolLM2-1.7B-Instruct-q4f16_1-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "1.7B",
-    quantization: "q0f16",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
-    },
-  },
-  {
-    name: "SmolLM-1.7B-Instruct-q0f32-MLC",
-    display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "1.7B",
-    quantization: "q0f32",
-    family: "SmolLM",
-    recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
-    },
-  },
-  {
-    name: "SmolLM-1.7B-Instruct-q4f16_1-MLC",
-    display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "1.7B",
-    quantization: "q4f16_1",
-    family: "SmolLM",
-    recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
-    },
-  },
-  {
-    name: "SmolLM-1.7B-Instruct-q4f32_1-MLC",
-    display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "1.7B",
-    quantization: "q4f32_1",
-    family: "SmolLM",
-    recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
-    },
-  },
-  {
-    name: "SmolLM-360M-Instruct-q0f16-MLC",
-    display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "360M",
-    quantization: "q0f16",
-    family: "SmolLM",
-    recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-360M-Instruct-q0f32-MLC",
+    name: "SmolLM2-1.7B-Instruct-q4f32_1-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "360M",
-    quantization: "q0f32",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-360M-Instruct-q4f16_1-MLC",
+    name: "SmolLM2-360M-Instruct-q0f16-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "360M",
-    quantization: "q4f16_1",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-360M-Instruct-q4f32_1-MLC",
+    name: "SmolLM2-360M-Instruct-q0f32-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "360M",
-    quantization: "q4f32_1",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-135M-Instruct-q0f16-MLC",
+    name: "SmolLM2-360M-Instruct-q4f16_1-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "135M",
-    quantization: "q0f16",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-135M-Instruct-q0f32-MLC",
+    name: "SmolLM2-360M-Instruct-q4f32_1-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "135M",
-    quantization: "q0f32",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-135M-Instruct-q4f16_1-MLC",
+    name: "SmolLM2-135M-Instruct-q0f16-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "135M",
-    quantization: "q4f16_1",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
-    name: "SmolLM-135M-Instruct-q4f32_1-MLC",
+    name: "SmolLM2-135M-Instruct-q0f32-MLC",
     display_name: "SmolLM",
-    provider: "SmolLM",
-    size: "135M",
-    quantization: "q4f32_1",
-    family: "SmolLM",
+    provider: "HuggingFaceTB",
+    family: "SmolLM2",
     recommended_config: {
-      temperature: 0.6,
-      top_p: 0.92,
+      temperature: 1,
+      presence_penalty: 0,
+      frequency_penalty: 0,
+      top_p: 1,
     },
   },
   {
     name: "Qwen2.5-0.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -594,8 +499,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-0.5B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -608,8 +511,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-0.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -622,8 +523,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-0.5B-Instruct-q0f16-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q0f16",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -636,8 +535,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-0.5B-Instruct-q0f32-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q0f32",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -650,8 +547,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-1.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -664,8 +559,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-1.5B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -678,8 +571,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-3B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "3B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -692,8 +583,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-3B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "3B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -706,8 +595,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-7B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -720,8 +607,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-7B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5",
     recommended_config: {
       temperature: 0.7,
@@ -735,8 +620,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-Coder-1.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5 Coder",
     recommended_config: {
       temperature: 1.0,
@@ -749,8 +632,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-Coder-1.5B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5 Coder",
     recommended_config: {
       temperature: 1.0,
@@ -763,8 +644,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-Coder-7B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Qwen 2.5 Coder",
     recommended_config: {
       temperature: 1.0,
@@ -777,8 +656,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2.5-Coder-7B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Qwen 2.5 Coder",
     recommended_config: {
       temperature: 1.0,
@@ -792,8 +669,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-Math-1.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2 Math",
     recommended_config: {
       temperature: 1.0,
@@ -806,8 +681,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-Math-1.5B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f32_1",
     family: "Qwen 2 Math",
     recommended_config: {
       temperature: 1.0,
@@ -820,8 +693,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-Math-7B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Qwen 2 Math",
     recommended_config: {
       temperature: 0.7,
@@ -834,8 +705,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-Math-7B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Qwen 2 Math",
     recommended_config: {
       temperature: 0.7,
@@ -848,8 +717,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-it-q4f16_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f16_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -862,8 +729,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-it-q4f32_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f32_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -876,8 +741,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-it-q4f16_1-MLC-1k",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f16_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -890,8 +753,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-it-q4f32_1-MLC-1k",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f32_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -904,8 +765,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-9b-it-q4f16_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "9B",
-    quantization: "q4f16_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -918,8 +777,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-9b-it-q4f32_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "9B",
-    quantization: "q4f32_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -932,8 +789,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-jpn-it-q4f16_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f16_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -946,8 +801,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "gemma-2-2b-jpn-it-q4f32_1-MLC",
     display_name: "Gemma",
     provider: "Google",
-    size: "2B",
-    quantization: "q4f32_1",
     family: "Gemma",
     recommended_config: {
       temperature: 0.7,
@@ -960,8 +813,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "stablelm-2-zephyr-1_6b-q4f16_1-MLC",
     display_name: "StableLM",
     provider: "Hugging Face",
-    size: "1.6B",
-    quantization: "q4f16_1",
     family: "StableLM 2",
     recommended_config: {
       temperature: 0.7,
@@ -974,8 +825,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "stablelm-2-zephyr-1_6b-q4f32_1-MLC",
     display_name: "StableLM",
     provider: "Hugging Face",
-    size: "1.6B",
-    quantization: "q4f32_1",
     family: "StableLM 2",
     recommended_config: {
       temperature: 0.7,
@@ -988,8 +837,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "stablelm-2-zephyr-1_6b-q4f16_1-MLC-1k",
     display_name: "StableLM",
     provider: "Hugging Face",
-    size: "1.6B",
-    quantization: "q4f16_1",
     family: "StableLM 2",
     recommended_config: {
       temperature: 0.7,
@@ -1002,8 +849,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "stablelm-2-zephyr-1_6b-q4f32_1-MLC-1k",
     display_name: "StableLM",
     provider: "Hugging Face",
-    size: "1.6B",
-    quantization: "q4f32_1",
     family: "StableLM 2",
     recommended_config: {
       temperature: 0.7,
@@ -1016,8 +861,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "RedPajama-INCITE-Chat-3B-v1-q4f16_1-MLC",
     display_name: "RedPajama",
     provider: "Together",
-    size: "3B",
-    quantization: "q4f16_1",
     family: "RedPajama",
     recommended_config: {
       temperature: 0.7,
@@ -1028,8 +871,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "RedPajama-INCITE-Chat-3B-v1-q4f32_1-MLC",
     display_name: "RedPajama",
     provider: "Together",
-    size: "3B",
-    quantization: "q4f32_1",
     family: "RedPajama",
     recommended_config: {
       temperature: 0.7,
@@ -1040,8 +881,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "RedPajama-INCITE-Chat-3B-v1-q4f16_1-MLC-1k",
     display_name: "RedPajama",
     provider: "Together",
-    size: "3B",
-    quantization: "q4f16_1",
     family: "RedPajama",
     recommended_config: {
       temperature: 0.7,
@@ -1052,8 +891,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "RedPajama-INCITE-Chat-3B-v1-q4f32_1-MLC-1k",
     display_name: "RedPajama",
     provider: "Together",
-    size: "3B",
-    quantization: "q4f32_1",
     family: "RedPajama",
     recommended_config: {
       temperature: 0.7,
@@ -1064,8 +901,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v1.0-q4f16_1-MLC",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f16_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 1,
@@ -1078,8 +913,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v1.0-q4f32_1-MLC",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f32_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 1,
@@ -1092,8 +925,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v1.0-q4f16_1-MLC-1k",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f16_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 1,
@@ -1106,8 +937,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v1.0-q4f32_1-MLC-1k",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f32_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 1,
@@ -1120,8 +949,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3.1-70B-Instruct-q3f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "70B",
-    quantization: "q3f16_1",
     family: "Llama 3.1",
     recommended_config: {
       temperature: 0.6,
@@ -1134,8 +961,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-0.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1148,8 +973,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-0.5B-Instruct-q0f16-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q0f16",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1162,8 +985,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-0.5B-Instruct-q0f32-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "0.5B",
-    quantization: "q0f32",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1176,8 +997,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-1.5B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f16_1",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1190,8 +1009,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-1.5B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "1.5B",
-    quantization: "q4f32_1",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1204,8 +1021,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-7B-Instruct-q4f16_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1218,8 +1033,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Qwen2-7B-Instruct-q4f32_1-MLC",
     display_name: "Qwen",
     provider: "Alibaba",
-    size: "7B",
-    quantization: "q4f32",
     family: "Qwen 2",
     recommended_config: {
       temperature: 0.7,
@@ -1232,8 +1045,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3-8B-Instruct-q4f32_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Llama 3",
     recommended_config: {
       temperature: 0.6,
@@ -1246,8 +1057,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3-8B-Instruct-q4f16_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Llama 3",
     recommended_config: {
       temperature: 0.6,
@@ -1260,8 +1069,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3-8B-Instruct-q4f32_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f32_1",
     family: "Llama 3",
     recommended_config: {
       temperature: 0.6,
@@ -1274,8 +1081,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3-8B-Instruct-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "8B",
-    quantization: "q4f16_1",
     family: "Llama 3",
     recommended_config: {
       temperature: 0.6,
@@ -1288,8 +1093,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-3-70B-Instruct-q3f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "70B",
-    quantization: "q3f16_1",
     family: "Llama 3",
     recommended_config: {
       temperature: 0.7,
@@ -1303,7 +1106,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3-mini-4k-instruct-q4f16_1-MLC",
     display_name: "Phi 3",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 3 Mini Instruct",
     recommended_config: {
       temperature: 0.7,
@@ -1316,7 +1118,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3-mini-4k-instruct-q4f32_1-MLC",
     display_name: "Phi 3",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 3 Mini Instruct",
     recommended_config: {
       temperature: 0.7,
@@ -1329,7 +1130,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3-mini-4k-instruct-q4f16_1-MLC-1k",
     display_name: "Phi 3",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 3 Mini Instruct",
     recommended_config: {
       temperature: 0.7,
@@ -1342,7 +1142,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Phi-3-mini-4k-instruct-q4f32_1-MLC-1k",
     display_name: "Phi 3",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 3 Mini Instruct",
     recommended_config: {
       temperature: 0.7,
@@ -1355,8 +1154,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-2-7b-chat-hf-q4f32_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Llama 2",
     recommended_config: {
       temperature: 0.6,
@@ -1367,8 +1164,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-2-7b-chat-hf-q4f16_1-MLC-1k",
     display_name: "Llama",
     provider: "Meta",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Llama 2",
     recommended_config: {
       temperature: 0.6,
@@ -1379,8 +1174,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-2-7b-chat-hf-q4f32_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "7B",
-    quantization: "q4f32_1",
     family: "Llama 2",
     recommended_config: {
       temperature: 0.6,
@@ -1391,8 +1184,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-2-7b-chat-hf-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "7B",
-    quantization: "q4f16_1",
     family: "Llama 2",
     recommended_config: {
       temperature: 0.6,
@@ -1403,8 +1194,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "Llama-2-13b-chat-hf-q4f16_1-MLC",
     display_name: "Llama",
     provider: "Meta",
-    size: "13B",
-    quantization: "q4f16_1",
     family: "Llama 2",
     recommended_config: {
       temperature: 0.6,
@@ -1415,7 +1204,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-2-q4f16_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 2",
     recommended_config: {
       temperature: 0.7,
@@ -1426,7 +1214,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-2-q4f32_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 2",
     recommended_config: {
       temperature: 0.7,
@@ -1437,7 +1224,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-2-q4f16_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 2",
     recommended_config: {
       temperature: 0.7,
@@ -1448,7 +1234,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-2-q4f32_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 2",
     recommended_config: {
       temperature: 0.7,
@@ -1459,7 +1244,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-1_5-q4f16_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 1.5",
     recommended_config: {
       temperature: 0.7,
@@ -1470,7 +1254,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-1_5-q4f32_1-MLC",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 1.5",
     recommended_config: {
       temperature: 0.7,
@@ -1481,7 +1264,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-1_5-q4f16_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f16_1",
     family: "Phi 1.5",
     recommended_config: {
       temperature: 0.7,
@@ -1492,7 +1274,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "phi-1_5-q4f32_1-MLC-1k",
     display_name: "Phi",
     provider: "Microsoft",
-    quantization: "q4f32_1",
     family: "Phi 1.5",
     recommended_config: {
       temperature: 0.7,
@@ -1503,8 +1284,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v0.4-q4f16_1-MLC",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f16_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 0.7,
@@ -1515,8 +1294,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v0.4-q4f32_1-MLC",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f32_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 0.7,
@@ -1527,8 +1304,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v0.4-q4f16_1-MLC-1k",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f16_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 0.7,
@@ -1539,8 +1314,6 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     name: "TinyLlama-1.1B-Chat-v0.4-q4f32_1-MLC-1k",
     display_name: "TinyLlama",
     provider: "Zhang Peiyuan",
-    size: "1.1B",
-    quantization: "q4f32_1",
     family: "TinyLlama",
     recommended_config: {
       temperature: 0.7,
@@ -1549,6 +1322,24 @@ export const DEFAULT_MODELS: ModelRecord[] = [
   },
 ];
 
+export const DEFAULT_MODELS: ModelRecord[] = DEFAULT_MODEL_BASES.filter(
+  (model) => {
+    if (
+      !prebuiltAppConfig.model_list.map((m) => m.model_id).includes(model.name)
+    ) {
+      console.warn(
+        `Model ${model.name} not supported by current WebLLM version.`,
+      );
+      return false;
+    }
+    return true;
+  },
+).map((model) => ({
+  ...model,
+  size: getSize(model.name),
+  quantization: getQuantization(model.name),
+}));
+
 export const CHAT_PAGE_SIZE = 15;
 export const MAX_RENDER_MSG_COUNT = 45;
 
diff --git a/app/store/config.ts b/app/store/config.ts
index 3f386e78..26ad6feb 100644
--- a/app/store/config.ts
+++ b/app/store/config.ts
@@ -217,9 +217,9 @@ export const useAppConfig = createPersistStore(
   }),
   {
     name: StoreKey.Config,
-    version: 0.59,
+    version: 0.6,
     migrate: (persistedState, version) => {
-      if (version < 0.59) {
+      if (version < 0.6) {
         return {
           ...DEFAULT_CONFIG,
           ...(persistedState as any),
diff --git a/app/utils.ts b/app/utils.ts
index 04b24cd8..4d184369 100644
--- a/app/utils.ts
+++ b/app/utils.ts
@@ -279,3 +279,23 @@ export function fixMessage(message: string) {
 
   return message;
 }
+
+// Get model size from model id
+export function getSize(model_id: string): string | undefined {
+  const sizeRegex = /-(\d+(\.\d+)?[BK])-?/;
+  const match = model_id.match(sizeRegex);
+  if (match) {
+    return match[1];
+  }
+  return undefined;
+}
+
+// Get quantization method from model id
+export function getQuantization(model_id: string): string | undefined {
+  const quantizationRegex = /-(q[0-9]f[0-9]+(?:_[0-9])?)-/;
+  const match = model_id.match(quantizationRegex);
+  if (match) {
+    return match[1];
+  }
+  return undefined;
+}
diff --git a/package.json b/package.json
index d2cf025e..615d462d 100644
--- a/package.json
+++ b/package.json
@@ -17,7 +17,7 @@
   "dependencies": {
     "@fortaine/fetch-event-source": "^3.0.6",
     "@hello-pangea/dnd": "^16.5.0",
-    "@mlc-ai/web-llm": "^0.2.73",
+    "@mlc-ai/web-llm": "^0.2.74",
     "@serwist/next": "^9.0.2",
     "@svgr/webpack": "^6.5.1",
     "emoji-picker-react": "^4.9.2",
diff --git a/yarn.lock b/yarn.lock
index df40e3d2..409eb1ff 100644
--- a/yarn.lock
+++ b/yarn.lock
@@ -1180,10 +1180,10 @@
     "@jridgewell/resolve-uri" "^3.1.0"
     "@jridgewell/sourcemap-codec" "^1.4.14"
 
-"@mlc-ai/web-llm@^0.2.73":
-  version "0.2.73"
-  resolved "https://registry.yarnpkg.com/@mlc-ai/web-llm/-/web-llm-0.2.73.tgz#42c208cac7f5efdd0df4ad5f974101437a094e49"
-  integrity sha512-v9eS05aptZYNBhRZ7KEq5dr1kMESHvkdwru9tNY1oS1GDDEBjUt4Y/u06ejvP3qF4VatCakEezzJT3SkuMhDnQ==
+"@mlc-ai/web-llm@^0.2.74":
+  version "0.2.74"
+  resolved "https://registry.yarnpkg.com/@mlc-ai/web-llm/-/web-llm-0.2.74.tgz#ad98e3661d8bc1e538f4bb0ca5a2f1c67cc4ea46"
+  integrity sha512-DqqQNrkHmfGOuuT0ZDRKxtRrU6CpcmjtBwn7QJDMRi0OyY357rEjiswf/9jxVS8AuGPdTIRwQsC3jLU8JxXE9w==
   dependencies:
     loglevel "^1.9.1"