@infernetprotocol/engine 0.1.73 → 0.1.79
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/package.json +1 -1
- package/src/backends/ollama.js +6 -0
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@infernetprotocol/engine",
|
|
3
|
-
"version": "0.1.
|
|
3
|
+
"version": "0.1.79",
|
|
4
4
|
"description": "Inference engine abstraction for Infernet — pluggable backends (Mojo+MAX, in-process stub) over a stable async-iterator API.",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"type": "module",
|
package/src/backends/ollama.js
CHANGED
|
@@ -95,6 +95,12 @@ export async function createOllamaBackend({
|
|
|
95
95
|
model: resolvedModel,
|
|
96
96
|
messages: Array.isArray(messages) ? messages : [],
|
|
97
97
|
stream: true,
|
|
98
|
+
// Qwen 3.5 and other reasoning models can spend the
|
|
99
|
+
// entire num_predict budget in their private thinking
|
|
100
|
+
// channel, leaving message.content empty. Infernet's
|
|
101
|
+
// protocol carries the user-visible answer, not chain
|
|
102
|
+
// of thought, so request a normal answer stream.
|
|
103
|
+
think: false,
|
|
98
104
|
options: {}
|
|
99
105
|
};
|
|
100
106
|
if (typeof temperature === "number") body.options.temperature = temperature;
|