Minor fix on the script (#106)

mlc-ai · May 22, 2023 · e1e2e0a · e1e2e0a
1 parent 911ea9b
commit e1e2e0a
Show file tree

Hide file tree

Showing 5 changed files with 7 additions and 8 deletions.
diff --git a/mlc-llm b/mlc-llm
diff --git a/scripts/gh_deploy_site.sh b/scripts/gh_deploy_site.sh
@@ -1,7 +1,7 @@
 #!/bin/bash
 set -euxo pipefail
 
-scripts/build_site.sh web/global_config.json
+scripts/build_site.sh web/gh-page-config.json
 
 git fetch
 git checkout -B gh-pages origin/gh-pages

diff --git a/scripts/local_deploy_site.sh b/scripts/local_deploy_site.sh
@@ -6,27 +6,26 @@ MLC_LLM_HOME_SET="${MLC_LLM_HOME:-}"
 if [ -z ${MLC_LLM_HOME_SET} ]; then
     export MLC_LLM_HOME="${MLC_LLM_HOME:-mlc-llm}"
 fi
-
 scripts/build_site.sh web/local-config.json
 
 echo "symlink parameter location to site.."
 
 if [ -d "$MLC_LLM_HOME/dist/vicuna-v1-7b-q4f32_0/params" ]; then
     rm -rf site/_site/dist/vicuna-v1-7b-q4f32_0
     mkdir -p site/_site/dist/vicuna-v1-7b-q4f32_0
-    ln -s $MLC_LLM_HOME/dist/vicuna-v1-7b-q4f32_0/params site/_site/dist/vicuna-v1-7b-q4f32_0/params
+    ln -s  "$(cd $MLC_LLM_HOME/dist/vicuna-v1-7b-q4f32_0/params && pwd)" site/_site/dist/vicuna-v1-7b-q4f32_0/params
     cp -rf $MLC_LLM_HOME/dist/vicuna-v1-7b-q4f32_0/vicuna-v1-7b-q4f32_0-webgpu.wasm site/_site/dist/vicuna-v1-7b-q4f32_0/
 fi
 if [ -d "$MLC_LLM_HOME/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/params" ]; then
     rm -rf site/_site/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0
     mkdir -p site/_site/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0
-    ln -s $MLC_LLM_HOME/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/params site/_site/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/params
+    ln -s "$(cd $MLC_LLM_HOME/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/params && pwd)" site/_site/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/params
     cp -rf $MLC_LLM_HOME/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/RedPajama-INCITE-Chat-3B-v1-q4f32_0-webgpu.wasm site/_site/dist/RedPajama-INCITE-Chat-3B-v1-q4f32_0/
 fi
 if [ -d "$MLC_LLM_HOME/dist/wizardlm-7b/params" ]; then
     rm -rf site/_site/dist/wizardlm-7b
     mkdir -p site/_site/dist/wizardlm-7b
-    ln -s $MLC_LLM_HOME/dist/wizardlm-7b/params site/_site/dist/wizardlm-7b/params
+    ln -s "$(cd $MLC_LLM_HOME/dist/wizardlm-7b/params && pwd)" site/_site/dist/wizardlm-7b/params
     cp -rf $MLC_LLM_HOME/dist/wizardlm-7b/wizardlm-7b-webgpu.wasm site/_site/dist/wizardlm-7b/
 fi
 

diff --git a/site/index.md b/site/index.md
@@ -37,7 +37,7 @@ If you have a Mac computer with Apple silicon, here are the instructions for you
 
 ## Chat Demo
 
-The chat demo is based on [vicuna-7b-v1.1](https://huggingface.co/lmsys/vicuna-7b-delta-v1.1) model. More model support are on the way.
+The chat demo is based on [vicuna-7b-v1.1](https://huggingface.co/lmsys/vicuna-7b-delta-v1.1) model and [RedPajama-INCITE-Chat-3B-v1](https://huggingface.co/togethercomputer/RedPajama-INCITE-Chat-3B-v1) model . More model support are on the way.
 
 {% include llm_chat.html %}
 

diff --git a/web/llm_chat.js b/web/llm_chat.js
@@ -187,7 +187,7 @@ class LLMChatPipeline {
       this.vm.getFunction("prefill")
     );
     this.decoding = this.tvm.detachFromCurrentScope(
-      this.vm.getFunction("decoding")
+      this.vm.getFunction("decode")
     );
     this.params = this.tvm.detachFromCurrentScope(
       this.tvm.getParamsFromCache("param", cacheMetadata.ParamSize)
+1 −1		3rdparty/tokenizers-cpp
+13 −9		README.md
+2 −2		android/MLCChat/app/src/main/java/ai/mlc/mlcchat/ChatState.java
+5 −5		android/MLCChat/app/src/main/java/ai/mlc/mlcchat/LLMChat.java
+37 −15		build.py
+144 −69		cpp/cli_main.cc
+23 −16		cpp/llm_chat.cc
+3 −3		ios/MLCChat/ChatState.swift
+42 −75		ios/MLCChat/LLMChat.mm
+70 −6		ios/MLCChat/MLCChat-Bridging-Header.h
+36 −36		log_db/dolly-v2-3b/database_tuning_record.json
+36 −36		log_db/dolly-v2-3b/database_workload.json
+35 −35		log_db/redpajama-3b-q4f16/database_tuning_record.json
+35 −35		log_db/redpajama-3b-q4f16/database_workload.json
+30 −30		log_db/redpajama-3b-q4f32/database_tuning_record.json
+30 −30		log_db/redpajama-3b-q4f32/database_workload.json
+18 −18		log_db/vicuna-v1-7b/database_tuning_record.json
+23 −23		log_db/vicuna-v1-7b/database_workload.json
+516 −2		mlc_llm/dispatch/dispatch_tir_operator_adreno.py
+1 −1		mlc_llm/dispatch/gpt_neox/dolly_v2_3b.py
+258 −4		mlc_llm/dispatch/llama/main.py
+4 −4		mlc_llm/relax_model/gpt_neox.py
+4 −4		mlc_llm/relax_model/llama.py
+4 −4		mlc_llm/relax_model/moss.py
+24 −6		mlc_llm/utils.py
+25 −17		site/index.md
+2 −2		tests/chat.py
+2 −2		tests/debug/compare_lib.py
+6 −6		tests/evaluate.py