Skip to content

yaml-time: v3 config update #173

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
SearchSavior wants to merge 5 commits into
base: main
Choose a base branch
from
Open

yaml-time: v3 config update #173

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
a81c61b
new around here? *prepare yourself, its yaml-time*
SearchSavior Sep 17, 2026
5d08e45
rewrite openarc add to support the new yaml config keys
SearchSavior Sep 18, 2026
6386f75
deprecate openarc add
SearchSavior Sep 19, 2026
bd8645c
updates to the config docs
SearchSavior Sep 19, 2026
2750a01
update docs for openarc configuration
SearchSavior Sep 19, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
271 changes: 163 additions & 108 deletions config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
server:
port:

port:

# comma seperated list of model_name to load on startup

Expand All @@ -9,16 +8,21 @@ startup_models : ${OPENRARC_AUTOLOAD_MODELS}


models:
# Each model entry nests its load-time fields under `load_config`. Model-level
# request defaults (sampler_config and the per-model_type *_config blocks) are
# siblings of load_config and are overridden per request:
# request-time > config.yaml > engine default
qwen35-08b:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen3.5/Qwen3.5-0.8B-int8_asym-ov/
device: CPU
tool_call_parser: qwen35
load_config:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen3.5/Qwen3.5-0.8B-int8_asym-ov/
device: CPU
tool_call_parser: qwen35
runtime_config:
PERFORMANCE_HINT: LATENCY
scheduler_config:
max_num_batched_tokens:
max_num_batched_tokens:
num_kv_blocks:
cache_size:
num_linear_attention_blocks:
Expand All @@ -28,7 +32,7 @@ models:
use_cache_eviction:
use_sparse_attention:
sampler_config:
temperature: 0.7
tempeature: 0.7
top_k: 40
top_p: 0.95
repetition_penalty: 1.05
Expand All @@ -37,7 +41,7 @@ models:
qwen3-8b-spec:
load_config:
engine: ovgenai
model_type: vlm
model_type: llm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen/Qwen3-8B-ShiningValiant3-int4-asym-ov
device: GPU.0
tool_call_parser: hermes
Expand All @@ -46,115 +50,166 @@ models:
num_assistant_tokens: 5
runtime_config:
PERFORMANCE_HINT: LATENCY
scheduler_config:
max_num_batched_tokens:
num_kv_blocks:
cache_size:
num_linear_attention_blocks:
cache_interval_multiplier:
dynamic_split_fuse:
enable_prefix_caching:
use_cache_eviction:
use_sparse_attention:
sampler_config:
temperature: 0.7
top_k: 40
top_p: 0.95
repetition_penalty: 1.05
max_tokens: 1024
#######################################################################################################
#######################################################################################################
#######################################################################################################
orion-v1:
load_config:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Orion-26B-A4B-v1-int4-ov/
device: GPU.1
tool_call_parser: gemma4
runtime_config:
ATTENTION_BACKEND: PA
CACHE_DIR: /mnt/Ironwolf-4TB/Models/OpenVINO/Orion-26B-A4B-v1-int4-ov/model_cache
#######################################################################################################
#######################################################################################################
#######################################################################################################
ornith-9b:
load_config:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen/Ornith-1.5-9B/Ornith-1.5-9B-int4_asym-awq-se256-ov
device: GPU.0
runtime_config:
ATTENTION_BACKEND: PA
CACHE_DIR: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen/Ornith-1.5-9B/Ornith-1.5-9B-int4_asym-awq-se256-ov/model_cache
# ====
# LLM
# qwen38-27b-davidau
#######################################################################################################
qwen38-27b-davidau:
load_config:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Huihui-Qwen3.8-27B-abliterated-int4-awq
device: GPU.0
tool_call_parser: qwen35
runtime_config:
ATTENTION_BACKEND: PA
CACHE_DIR: /mnt/Ironwolf-4TB/Models/OpenVINO/Huihui-Qwen3.8-27B-abliterated-int4-awq/model_cache
scheduler_config:
cache_interval_multiplier: 128
max_num_batched_tokens: 2048
enable_prefix_caching: true
use_sparse_attention: true
max_num_seqs: 1
num_linear_attention_blocks: 0
# ===========================
# LLMs
# ===========================
qwen38-27b:
load_config:
engine: ovgenai
model_type: vlm
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen3.5/Qwen3.8-27B-int4-ov/
device: GPU.0
tool_call_parser: qwen35
runtime_config:
ATTENTION_BACKEND: PA
CACHE_DIR: /mnt/Ironwolf-4TB/Models/OpenVINO/Qwen3.5/Qwen3.8-27B-int4-ov/model_cache
scheduler_config:
cache_interval_multiplier: 128
max_num_batched_tokens: 2048
enable_prefix_caching: true
use_sparse_attention: true
max_num_seqs: 1
num_linear_attention_blocks: 0

kokoro:
engine: openvino
model_type: kokoro
model_path:
device: CPU
load_config:
engine: openvino
model_type: kokoro
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Kokoro-82M-FP16-OpenVINO
device: CPU
kokoro_config:
voice: af_sarah
voice_blend: af_heart:0.7,af_nicole:0.3
lang_code:
speed: 1.0
chracter_count_chunk: 100
voice_blend: af_heart:0.7,af_nicole:0.3

qwen3asr:
engine: openvino
model_type: qwen3_asr
model_path:
device: CPU
qwen3_asr_config:
language:
max_tokens:
max_chunk_sec:
search_expand_sec:
min_window_ms:
whisper:
load_config:
engine: ovgenai
model_type: whisper
model_path: /mnt/Ironwolf-4TB/Models/OpenVINO/Whisper/distil-whisper-large-v3-int8-ov
device: GPU.0


qwen3_tts_elmo:
model_type: qwen3_tts_voice_clone
model_path:
engine: openvino
device: GPU.0
qwen3_tts_custom_voice_config:
language:
speaker:
instruct:
qwen3_tts_config:
max_new_tokens:
do_sample:
top_k:
top_p:
temperature:
repetition_penalty:
subtalker_do_sample:
subtalker_top_k:
subtalker_top_p:
subtalker_temperature:
stream:
stream_chunk_frames:
stream_left_context:
# --- Incomplete examples: fill in model_path to enable these. ---
# kokoro:
# load_config:
# engine: openvino
# model_type: kokoro
# model_path:
# device: CPU
# kokoro_config:
# voice: af_sarah
# voice_blend: af_heart:0.7,af_nicole:0.3
# lang_code:
# speed: 1.0
# character_count_chunk: 100

# qwen3asr:
# load_config:
# engine: openvino
# model_type: qwen3_asr
# model_path:
# device: CPU
# qwen3_asr_config:
# language:
# max_tokens:
# max_chunk_sec:
# search_expand_sec:
# min_window_ms:

qwen3_tts_oscar:
model_type: qwen3_tts_voice_design
model_path:
engine: openvino
device: GPU.0
qwen3_tts_voice_design_config:
voice_description:
qwen3_tts_config:
max_new_tokens:
do_sample:
top_k:
top_p:
temperature:
repetition_penalty:
subtalker_do_sample:
subtalker_top_k:
subtalker_top_p:
subtalker_temperature:
stream:
stream_chunk_frames:
stream_left_context:
# qwen3_tts_oscar:
# load_config:
# model_type: qwen3_tts_voice_design
# model_path:
# engine: openvino
# device: GPU.0
# qwen3_tts_voice_design_config:
# voice_description:
# qwen3_tts_config:
# max_new_tokens:
# do_sample:
# top_k:
# top_p:
# temperature:
# repetition_penalty:
# subtalker_do_sample:
# subtalker_top_k:
# subtalker_top_p:
# subtalker_temperature:
# stream:
# stream_chunk_frames:
# stream_left_context:

qwen3_tts_oscar:
model_type: qwen3_tts_voice_clone
model_path:
engine: openvino
device: GPU.0
qwen3_tts_voice_clone_config:
ref_text:
x_vector_only:
instruct:
qwen3_tts_config:
max_new_tokens:
do_sample:
top_k:
top_p:
temperature:
repetition_penalty:
subtalker_do_sample:
subtalker_top_k:
subtalker_top_p:
subtalker_temperature:
stream:
stream_chunk_frames:
stream_left_context:
# qwen3_tts_elmo:
# load_config:
# model_type: qwen3_tts_voice_clone
# model_path:
# engine: openvino
# device: GPU.0
# qwen3_tts_voice_clone_config:
# ref_text:
# x_vector_only:
# instruct:
# qwen3_tts_config:
# max_new_tokens:
# do_sample:
# top_k:
# top_p:
# temperature:
# repetition_penalty:
# subtalker_do_sample:
# subtalker_top_k:
# subtalker_top_p:
# subtalker_temperature:
# stream:
# stream_chunk_frames:
# stream_left_context:
Loading
Loading