summaryrefslogtreecommitdiff
path: root/lib/hyperstack
diff options
context:
space:
mode:
authorPaul Buetow <paul@buetow.org>2026-05-24 15:09:01 +0300
committerPaul Buetow <paul@buetow.org>2026-05-24 15:09:01 +0300
commit897c9f99a8ae03c40f672b39bd8cadb2df9b1bdc (patch)
treea31878388612a15702efe806e03b558cc99d5a02 /lib/hyperstack
parentdffa66a331350fb7995bf93cff49b978a9288459 (diff)
chore(config): remove gpt-oss-120b references since qwen3.6 is better
Diffstat (limited to 'lib/hyperstack')
-rw-r--r--lib/hyperstack/config.rb4
-rw-r--r--lib/hyperstack/manager.rb2
-rw-r--r--lib/hyperstack/provisioning.rb2
3 files changed, 4 insertions, 4 deletions
diff --git a/lib/hyperstack/config.rb b/lib/hyperstack/config.rb
index 7057b4f..e41d1dd 100644
--- a/lib/hyperstack/config.rb
+++ b/lib/hyperstack/config.rb
@@ -49,7 +49,7 @@ module HyperstackVM
'assign_floating_ip' => true,
'create_bootable_volume' => false,
'enable_port_randomization' => false,
- 'labels' => %w[gpt-oss-120b wireguard]
+ 'labels' => %w[qwen36-27b wireguard]
},
'ssh' => {
'username' => 'ubuntu',
@@ -81,7 +81,7 @@ module HyperstackVM
'gpu_overhead_mb' => 2000,
'num_parallel' => 1,
'context_length' => 32_768,
- 'pull_models' => ['qwen3-coder:30b', 'gpt-oss:20b', 'gpt-oss:120b', 'nemotron-3-super']
+ 'pull_models' => ['qwen3-coder:30b', 'qwen36-27b', 'nemotron-3-super']
},
'vllm' => {
'install' => true,
diff --git a/lib/hyperstack/manager.rb b/lib/hyperstack/manager.rb
index 7a68199..e8382bb 100644
--- a/lib/hyperstack/manager.rb
+++ b/lib/hyperstack/manager.rb
@@ -938,7 +938,7 @@ module HyperstackVM
req.body = JSON.generate(
'model' => model,
'messages' => [{ 'role' => 'user', 'content' => prompt }],
- # 500 tokens: reasoning models (e.g. gpt-oss) use tokens for chain-of-thought
+ # 500 tokens: reasoning models use tokens for chain-of-thought
# before content; 50 is too small and yields an empty content field.
'max_tokens' => 500
)
diff --git a/lib/hyperstack/provisioning.rb b/lib/hyperstack/provisioning.rb
index 0b56559..948cd2a 100644
--- a/lib/hyperstack/provisioning.rb
+++ b/lib/hyperstack/provisioning.rb
@@ -143,7 +143,7 @@ module HyperstackVM
tp_size = Integer(cfg['tensor_parallel_size'] || @config.vllm_tensor_parallel_size)
parser = cfg['tool_call_parser']
# parser is nil only when preset explicitly omits the key and config has no default;
- # empty string means "disable tool calling" (e.g. gpt-oss reasoning models).
+ # empty string means "disable tool calling" (e.g. reasoning models).
parser = @config.vllm_tool_call_parser if parser.nil?
# Fall back to the top-level [vllm] config values when no preset is in use.
# This allows setting trust_remote_code / extra_vllm_args in the default [vllm] block