Views
No views yet
| Metric | Value |
|---|---|
| Model Size | ~4.38 GB (4-bit quantized) |
| Inference Speed | 30.58 tokens/sec (M1 MAX) |
| 112.80 tokens/sec (M3 ULTRA) | |
| gguf Q4_K_S | 8.14 tokens/sec (M1 MAX) |
| Context Support | 192,000 tokens |
1{%- if tools %}
2 {{- '\/system\n' }}
3 {%- if messages[0].role == 'system' %}
4 {{- messages[0].content + '\n\n' }}
5 {%- endif %}
6 {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
7 {%- for tool in tools %}
8 {{- "\n" }}
9 {{- tool | tojson }}
10 {%- endfor %}
11 {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call>...</tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n" }}
12{%- else %}
13 {%- if messages[0].role == 'system' %}
14 {{- '\/system\n' + messages[0].content + '\/\n' }}
15 {%- endif %}
16{%- endif %}
17
18{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
19{%- for message in messages[::-1] %}
20 {%- set index = (messages|length - 1) - loop.index0 %}
21 {%- set tool_start = "｟" %}
22 {%- set tool_start_length = tool_start|length %}
23 {%- set start_of_message = message.content[:tool_start_length] %}
24 {%- set tool_end = "｠" %}
25 {%- set tool_end_length = tool_end|length %}
26 {%- set start_pos = (message.content|length) - tool_end_length %}
27 {%- if start_pos < 0 %}
28 {%- set start_pos = 0 %}
29 {%- endif %}
30 {%- set end_of_message = message.content[start_pos:] %}
31 {%- if ns.multi_step_tool and message.role == "user" and not(start_of_message == tool_start and end_of_message == tool_end) %}
32 {%- set ns.multi_step_tool = false %}
33 {%- set ns.last_query_index = index %}
34 {%- endif %}
35{%- endfor %}
36
37{%- for message in messages %}
38 {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
39 {{- '\/' + message.role + '\n' + message.content + '\/' + '\n' }}
40 {%- elif message.role == "assistant" %}
41 {%- set content = message.content %}
42 {%- set reasoning_content = '' %}
43 {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
44 {%- set reasoning_content = message.reasoning_content %}
45 {%- else %}
46 {%- if '\/' in message.content %}
47 {%- set content = (message.content.split('\/')|last).lstrip('\n') %}
48 {%- set reasoning_content = (message.content.split('\/')|first).rstrip('\n') %}
49 {%- set reasoning_content = (reasoning_content.split('')|last).lstrip('\n') %}
50 {%- endif %}
51 {%- endif %}
52
53 {%- if loop.index0 > ns.last_query_index %}
54 {%- if loop.last or (not loop.last and reasoning_content) %}
55 {{- '\/' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\/\n' + content.lstrip('\n') }}
56 {%- else %}
57 {{- '\/' + message.role + '\n' + content }}
58 {%- endif %}
59 {%- else %}
60 {{- '\/' + message.role + '\n' + content }}
61 {%- endif %}
62
63 {%- if message.tool_calls %}
64 {%- for tool_call in message.tool_calls %}
65 {%- if (loop.first and content) or (not loop.first) %}
66 {{- '\n' }}
67 {%- endif %}
68 {%- if tool_call.function %}
69 {%- set tool_call = tool_call.function %}
70 {%- endif %}
71 {{- '<tool_call>\n{"name": "' }}
72 {{- tool_call.name }}
73 {{- '", "arguments": ' }}
74 {%- if tool_call.arguments is string %}
75 {{- tool_call.arguments }}
76 {%- else %}
77 {{- tool_call.arguments | tojson }}
78 {%- endif %}
79 {{- '}\n</tool_call>' }}
80 {%- endfor %}
81 {%- endif %}
82 {{- '\/\n' }}
83 {%- elif message.role == "tool" %}
84 {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
85 {{- '\/user' }}
86 {%- endif %}
87 {{- '\n｟\n' }}
88 {{- message.content }}
89 {{- '\n｠' }}
90 {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
91 {{- '\/\n' }}
92 {%- endif %}
93 {%- endif %}
94{%- endfor %}
95
96{%- if add_generation_prompt %}
97 {{- '\/assistant\n' }}
98 {%- if enable_thinking is defined and enable_thinking is false %}
99 {{- 'text-generationnot-for-all-audiences, conversational, mlx1# Install MLX (Apple Silicon only)
2pip install mlx
3
4# Load model with Hugging Face Transformers
5from transformers import AutoModelForCausalLM, AutoTokenizer
6
7model = AutoModelForCausalLM.from_pretrained("Goraint/Qwen3-8b-192k-Context-6X-Josiefied-Uncensored-MLX-AWQ-4bit", device_map="auto")
8tokenizer = AutoTokenizer.from_pretrained("Goraint/Qwen3-8b-192k-Context-6X-Josiefied-Uncensored-MLX-AWQ-4bit")1prompt = "Explain quantum computing in simple terms."
2inputs = tokenizer(prompt, return_tensors="pt").to("mps")
3outputs = model.generate(**inputs, max_length=200)
4print(tokenizer.decode(outputs[0], skip_special_tokens=True))Apache 2.0