v0.4.1 unstable work on getting full GPU usage and seamless playback (#…

…18)

* misc updates to 0.4.0

- reduce concurrency for alice
- adjustments to alice
- add context size to alice for gemma
- tune alice down
- configure all the variables
- comment out twitch handle for now
- setup running arc atomic bool for twitch thread
- thread stop's, default no subtitles
- concurrency \ spacing
- remove extra thread layer unneeded and potential deadlocking code.

* bump version to 0.4.1

FYI: unstable in some cases so far vs. 0.3.0

---------

Co-authored-by: Chris Kennedy <[email protected]>

Cargo.toml

@@ -17,7 +17,7 @@ license-file = "LICENSE"
 homepage = "https://github.com/groovybits/rsllm/wiki"
 repository = "https://github.com/groovybits/rsllm"
 authors = ["Chris Kennedy"]
-version = "0.4.0"
+version = "0.4.1"
 edition = "2021"
 
 [lib]

scripts/alice.sh

@@ -8,9 +8,16 @@
 #
 #
 
-BUILD_TYPE=debug
+BUILD_TYPE=release
 MODEL=mistral
 MODEL_ID=7b-it
+MAX_TOKENS=1000
+ALIGNMENT=right
+TEMPERATURE=0.8
+POLL_INTERVAL=10
+IMAGE_CONCURRENCY=1
+SPEECH_CONCURRENCY=1
+CONTEXT_SIZE=3000
 DYLD_LIBRARY_PATH=`pwd`:/usr/local/lib:$DYLD_LIBRARY_PATH \
     RUST_BACKTRACE=full target/${BUILD_TYPE}/rsllm \
     --query "create a story based on an anime About Alice an adult twitch streaming girl who lives in AI Wonderland. Have it vary off the title 'Alice in AI Wonderland' with a random plotline you create based on classic anime characters appearing in the wonderland. Alices AI Wonderland is a  happy fun show where Alice goes through experiences similar to Alice in Wonderland where she grows small or large depending one what she eats. Add in AI technology twists. Have it fully formatted like a transcript with the character speaking parts mostly speaking in first person, minimal narration. create a whole episode full length with classic anime characters with Alice the main character of AI Wonderland." \
@@ -22,10 +29,11 @@ DYLD_LIBRARY_PATH=`pwd`:/usr/local/lib:$DYLD_LIBRARY_PATH \
     --ndi-images \
     --mimic3-tts \
     --model-id $MODEL_ID \
-    --image-alignment right \
-    --temperature 0.8 \
-    --image-concurrency 1 \
-    --speech-concurrency 1\
-    --max-concurrent-sd-image-tasks 8 \
+    --image-alignment $ALIGNMENT \
+    --temperature $TEMPERATURE \
+    --image-concurrency $IMAGE_CONCURRENCY \
+    --speech-concurrency $SPEECH_CONCURRENCY \
+    --poll-interval $POLL_INTERVAL \
+    --llm-history-size $CONTEXT_SIZE \
     --daemon \
-    --max-tokens 1200 $@
+    --max-tokens $MAX_TOKENS $@

src/args.rs

@@ -4,7 +4,7 @@ use clap::Parser;
 #[derive(Parser, Debug, Clone)]
 #[clap(
     author = "Chris Kennedy",
-    version = "0.4.0",
+    version = "0.4.1",
     about = "Rust LLM Stream Analyzer and Content Generator"
 )]
 pub struct Args {
@@ -17,7 +17,7 @@ pub struct Args {
     )]
     pub system_prompt: String,
 
-    /// System prompt
+    /// Prompt
     #[clap(
         long,
         env = "QUERY",
@@ -596,6 +596,15 @@ pub struct Args {
     )]
     pub image_alignment: String,
 
+    /// Subtitles - enable subtitles
+    #[clap(
+        long,
+        env = "SUBTITLES",
+        default_value_t = false,
+        help = "Subtitles - enable subtitles."
+    )]
+    pub subtitles: bool,
+
     /// Subtitle position - top, mid-top, center, mid-bottom, bottom - bottom is default
     #[clap(
         long,