FEAT: Support QwQ-32B

[cyhasuka]

No description provided.

[ruifenggong]

建议加上 https://www.modelscope.cn/models/Qwen/QwQ-32B-GGUF

[cyhasuka]

建议加上 https://www.modelscope.cn/models/Qwen/QwQ-32B-GGUF

Added.

[qinxuye]

About MLX version, you could refer to codes below to make it more compact.

inference/xinference/model/llm/llm_family.json

Lines 1446 to 1456 in aeb1ccd

	{
	"model_format": "mlx",
	"model_size_in_billions": 70,
	"quantizations": [
	"3bit",
	"4bit",
	"6bit",
	"8bit",
	"fp16"
	],
	"model_id": "mlx-community/Llama-3.3-70B-Instruct-{quantization}"

[cyhasuka]

About MLX version, you could refer to codes below to make it more compact.

inference/xinference/model/llm/llm_family.json

Lines 1446 to 1456 in aeb1ccd

	{
	"model_format": "mlx",
	"model_size_in_billions": 70,
	"quantizations": [
	"3bit",
	"4bit",
	"6bit",
	"8bit",
	"fp16"
	],
	"model_id": "mlx-community/Llama-3.3-70B-Instruct-{quantization}"

Modified.

[qinxuye]

All tests failed, seems some places in your jsons have wrong syntax.

[qinxuye]

Don't know what has happened.

[cyhasuka]

uh, I use vscode instead of vim to look for json formatting issues, maybe the encoding is different

[qinxuye]

uh, I use vscode instead of vim to look for json formatting issues, maybe the encoding is different

OK, could you resolve it?

[cyhasuka]

OK, could you resolve it?

Solved.

[qinxuye]

LGTM

[ascacl]

你好，我容器化部署的，我进入容器按照这个代码修改了代码，然后重启了容器，为什么界面上选不到呢？

[oubeichen]

修改容器之后 commit 成新镜像了吗？不然不是自动还原了？

[ascacl]

修改容器之后 commit 成新镜像了吗？不然不是自动还原了？

docker exec进入容器，修改代码后，docker restart容器，这样不行吗？

[helojo]

Only ['qwen1.5-chat', 'qwen1.5-moe-chat', 'qwen2-instruct', 'qwen2-moe-instruct', 'qwen2.5-instruct', 'qwen2.5-coder-instruct', 'glm4-chat', 'glm4-chat-1m', 'llama-3.1-instruct', 'deepseek-r1-distill-qwen', 'deepseek-r1-distill-llama'] support tool calls。
could QwQ 32B enable function calls？