feat: upgrade llama-rs to llm (#62)

* feat: upgrade llama-rs to llm

Author: hlhr202

.github/workflows/llama-build.yml

@@ -68,7 +68,7 @@ jobs:
       - run: git submodule update --init --recursive
       - run: pnpm install
       - run: cargo clippy --all-targets
-      - run: pnpm build:llama-rs
+      - run: pnpm build:llm-rs
       - run: pnpm build:llama-cpp
       - run: pnpm build:rwkv-cpp
       - run: pnpm build

Cargo.lock

@@ -36,7 +36,7 @@ llama-node: Node.js Library for Large Language Model
 
 This project is in an early stage, the API for nodejs may change in the future, use it with caution.
 
-This is a nodejs library for inferencing llama, rwkv or llama derived models. It was built on top of [llama-rs](https://github.com/rustformers/llama-rs), [llama.cpp](https://github.com/ggerganov/llama.cpp) and [rwkv.cpp](https://github.com/saharNooby/rwkv.cpp). It uses [napi-rs](https://github.com/napi-rs/napi-rs) for channel messages between node.js and llama thread.
+This is a nodejs library for inferencing llama, rwkv or llama derived models. It was built on top of [llm (originally llama-rs)](https://github.com/rustformers/llm), [llama.cpp](https://github.com/ggerganov/llama.cpp) and [rwkv.cpp](https://github.com/saharNooby/rwkv.cpp). It uses [napi-rs](https://github.com/napi-rs/napi-rs) for channel messages between node.js and llama thread.
 
 Currently supported models (All of these should be converted to [GGML](https://github.com/ggerganov/ggml) format):
 
@@ -77,7 +77,7 @@ npm install llama-node
     npm install @llama-node/llama-cpp
     ```
 
-    -   or llama-rs
+    -   or llm
 
     ```bash
     npm install @llama-node/core
@@ -106,7 +106,7 @@ This library was published under MIT/Apache-2.0 license. However, we strongly re
 -   LLaMA models: [facebookresearch/llama](https://github.com/facebookresearch/llama)
 -   RWKV models: [BlinkDL/RWKV-LM](https://github.com/BlinkDL/RWKV-LM)
 -   llama.cpp: [ggreganov/llama.cpp](https://github.com/ggerganov/llama.cpp)
--   llama-rs: [rustformers/llama-rs](https://github.com/rustformers/llama-rs)
+-   llm: [rustformers/llm](https://github.com/rustformers/llm)
 -   rwkv.cpp: [saharNooby/rwkv.cpp](https://github.com/saharNooby/rwkv.cpp)
 
 ### Some source code comes from

README.md

@@ -1,8 +1,8 @@
 import { LLM } from "llama-node";
-import { LLamaRS } from "llama-node/dist/llm/llama-rs.js";
+import { LLMRS } from "llama-node/dist/llm/llm-rs.js";
 import path from "path";
 const model = path.resolve(process.cwd(), "../ggml-alpaca-7b-q4.bin");
-const llama = new LLM(LLamaRS);
+const llama = new LLM(LLMRS);
 const template = `how are you`;
 const prompt = `Below is an instruction that describes a task. Write a response that appropriately completes the request.
 
@@ -14,7 +14,7 @@ ${template}
 const params = {
     prompt,
     numPredict: 128,
-    temp: 0.2,
+    temperature: 0.2,
     topP: 1,
     topK: 40,
     repeatPenalty: 1,
@@ -24,7 +24,7 @@ const params = {
 };
 const run = async () => {
     const abortController = new AbortController();
-    await llama.load({ path: model });
+    await llama.load({ modelPath: model, modelType: "Llama" /* ModelType.Llama */ });
     setTimeout(() => {
         abortController.abort();
     }, 3000);

example/js/llama-rs/abortable.js example/js/llm-rs/abortable.jsexample/js/llama-rs/abortable.js renamed to example/js/llm-rs/abortable.js

@@ -1,14 +1,14 @@
 import { LLM } from "llama-node";
-import { LLamaRS } from "llama-node/dist/llm/llama-rs.js";
+import { LLMRS } from "llama-node/dist/llm/llm-rs.js";
 import path from "path";
 import fs from "fs";
 const model = path.resolve(process.cwd(), "../ggml-alpaca-7b-q4.bin");
-const llama = new LLM(LLamaRS);
+const llama = new LLM(LLMRS);
 const getWordEmbeddings = async (prompt, file) => {
     const data = await llama.getEmbedding({
         prompt,
         numPredict: 128,
-        temp: 0.2,
+        temperature: 0.2,
         topP: 1,
         topK: 40,
         repeatPenalty: 1,
@@ -19,7 +19,7 @@ const getWordEmbeddings = async (prompt, file) => {
     await fs.promises.writeFile(path.resolve(process.cwd(), file), JSON.stringify(data));
 };
 const run = async () => {
-    await llama.load({ path: model });
+    await llama.load({ modelPath: model, modelType: "Llama" /* ModelType.Llama */ });
     const dog1 = `My favourite animal is the dog`;
     await getWordEmbeddings(dog1, "./example/semantic-compare/dog1.json");
     const dog2 = `I have just adopted a cute dog`;

example/js/llama-rs/embedding.js example/js/llm-rs/embedding.jsexample/js/llama-rs/embedding.js renamed to example/js/llm-rs/embedding.js

@@ -1,8 +1,8 @@
 import { LLM } from "llama-node";
-import { LLamaRS } from "llama-node/dist/llm/llama-rs.js";
+import { LLMRS } from "llama-node/dist/llm/llm-rs.js";
 import path from "path";
 const model = path.resolve(process.cwd(), "../ggml-alpaca-7b-q4.bin");
-const llama = new LLM(LLamaRS);
+const llama = new LLM(LLMRS);
 const template = `how are you`;
 const prompt = `Below is an instruction that describes a task. Write a response that appropriately completes the request.
 
@@ -14,7 +14,7 @@ ${template}
 const params = {
     prompt,
     numPredict: 128,
-    temp: 0.2,
+    temperature: 0.2,
     topP: 1,
     topK: 40,
     repeatPenalty: 1,
@@ -23,7 +23,7 @@ const params = {
     feedPrompt: true,
 };
 const run = async () => {
-    await llama.load({ path: model });
+    await llama.load({ modelPath: model, modelType: "Llama" /* ModelType.Llama */ });
     await llama.createCompletion(params, (response) => {
         process.stdout.write(response.token);
     });