include llamafile repo as a submodule and make build process more self contained

.args

@@ -1,3 +1,2 @@
 -m
 TinyLLama-v0-5M-F16.gguf
-...

.gitmodules

@@ -4,3 +4,6 @@
 [submodule "llama.cpp"]
 	path = llama.cpp
 	url = [email protected]:mofosyne/llama.cpp.git
+[submodule "llamafile"]
+	path = llamafile
+	url = [email protected]:Mozilla-Ocho/llamafile.git

TinyLLama-v0-5M-F16.llamafile

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9847179c1f3a04f49deb90efa62173d02e10f272fe6cfe6325c6277cebd1b054
-size 17633471
+oid sha256:7f098bd53e7e689289be7322d19304eeea6f3b305ddbcfda4a15a452a1be35d1
+size 17633572

llama.cpp

@@ -1 +1 @@
-Subproject commit 8f4412980b41ccdc164ff220bfcd564f2a4a86cb
+Subproject commit da064a809badd5086d61fd82accbb16ad93cde94

llamafile

@@ -0,0 +1 @@
+Subproject commit cb92b32a6dfae4dff06d7333afe51f2b7224f709

llamafile-creation.sh

@@ -1,30 +1,46 @@
-#!/bin/sh
+#!/bin/bash
 
+###############################################################################
+echo == Prep Enviroment ==
 # Pull both the model folder and llama.cpp (for the conversion script)
-git submodule update --init
+#git submodule update --init
 
-# Convert from safetensor to gguf
-# (Assuming llama.cpp is in the next folder)
-./llama.cpp/convert.py maykeye_tinyllama --metadata maykeye_tinyllama-metadata.json
+###############################################################################
+echo == Build and prep the llamafile engine execuable ==
+pushd llamafile
+make -j8
+make
+# This is where each executables is located for reference purpose for now as of 2024-04-05
+# and was determined by running `sudo make install PREFIX=/usr/local`
+# ./o/llamafile/zipalign --> /usr/local/bin/zipalign
+# ./o/llama.cpp/main/main --> /usr/local/bin/llamafile
+# ./o/llama.cpp/imatrix/imatrix --> /usr/local/bin/llamafile-imatrix
+# ./o/llama.cpp/quantize/quantize --> /usr/local/bin/llamafile-quantize
+# ./build/llamafile-convert --> /usr/local/bin/llamafile-convert
+# ./o/llama.cpp/perplexity/perplexity --> /usr/local/bin/llamafile-perplexity
+# ./o/llama.cpp/llava/llava-quantize --> /usr/local/bin/llava-quantize
+popd
 
-# Copy the generated gguf to this folder
+###############################################################################
+echo == Convert from safetensor to gguf ==
+./llama.cpp/convert.py maykeye_tinyllama --outtype f16 --metadata maykeye_tinyllama-metadata.json
 mv maykeye_tinyllama/TinyLLama-v0-5M-F16.gguf TinyLLama-v0-5M-F16.gguf
 
-# Get the llamafile engine
-cp /usr/local/bin/llamafile TinyLLama-v0-5M-F16.llamafile
+echo == Generating Llamafile ==
+cp ./llamafile/o/llama.cpp/main/main TinyLLama-v0-5M-F16.llamafile
 
 # Create an .args file with settings defaults
 cat >.args <<EOF
 -m
 TinyLLama-v0-5M-F16.gguf
-...
 EOF
 
-# Combine
+# zip align engine, gguf and default args
 zipalign -j0 \
   TinyLLama-v0-5M-F16.llamafile \
   TinyLLama-v0-5M-F16.gguf \
   .args
 
-# Test
-./TinyLLama-v0-5M-F16.llamafile --cli -p "hello world the gruff man said"
+###############################################################################
+echo == Test Output ==
+./TinyLLama-v0-5M-F16.llamafile --cli -p "hello world the gruff man said"