Infinite thiking
#5
by KralHarlaus - opened
Using LMStudio default settings on Strix Halo machine Windows. after around 40k context, it starts thinking non stop and turns in to repeating the same sentence forever.
There is a proposed fix for the prompt template here, you may want to give it a change (and add some feedback if it helps). Click of the origin repo to get the full template.
Did the fix in the chat template helped?
can you share the prompt? (only if everything is public of course!)
@lgusm Hi, there. Since I started using Gemma3 26B-A4B I have been using the fixed template, and I do not recall having problems with endless repetitions, and it worked fine inside Zed with their builtin agent. Some remarks:
- I keep context up to 65536 (64k)
- I use q8_0 quantization for both the main model and draft KV caches
- Sampling parameters as recommended elsewhere:
repeat-penalty = 1.05
temp = 1.0
top-p = 0.95
top-k = 64
chat_template.jinja - from the link above, but not the latest commit (Click to Expand)
{%- macro format_parameters(properties, required, filter_keys=false) -%}
{%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
{%- set ns = namespace(found_first=false) -%}
{%- for key, value in properties | dictsort -%}
{%- set add_comma = false -%}
{%- if not filter_keys or key not in standard_keys -%}
{%- if ns.found_first %},{% endif -%}
{%- set ns.found_first = true -%}
{{ key }}:{
{%- if value['description'] -%}
description:<|"|>{{ value['description'] }}<|"|>
{%- set add_comma = true -%}
{%- endif -%}
{%- if value['type'] | upper == 'STRING' -%}
{%- if value['enum'] -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
enum:{{ format_argument(value['enum']) }}
{%- endif -%}
{%- elif value['type'] | upper == 'ARRAY' -%}
{%- if value['items'] is mapping and value['items'] -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
items:{
{%- set ns_items = namespace(found_first=false) -%}
{%- for item_key, item_value in value['items'] | dictsort -%}
{%- if item_value is not none -%}
{%- if ns_items.found_first %},{% endif -%}
{%- set ns_items.found_first = true -%}
{%- if item_key == 'properties' -%}
properties:{
{%- if item_value is mapping -%}
{{- format_parameters(item_value, value['items']['required'] | default([])) -}}
{%- endif -%}
}
{%- elif item_key == 'required' -%}
required:[
{%- for req_item in item_value -%}
<|"|>{{- req_item -}}<|"|>
{%- if not loop.last %},{% endif -%}
{%- endfor -%}
]
{%- elif item_key == 'type' -%}
{%- if item_value is string -%}
type:{{ format_argument(item_value | upper) }}
{%- else -%}
type:{{ format_argument(item_value | map('upper') | list) }}
{%- endif -%}
{%- else -%}
{{ item_key }}:{{ format_argument(item_value) }}
{%- endif -%}
{%- endif -%}
{%- endfor -%}
}
{%- endif -%}
{%- endif -%}
{%- if value['nullable'] %}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
nullable:true
{%- endif -%}
{%- if value['type'] | upper == 'OBJECT' -%}
{%- if value['properties'] is defined and value['properties'] is mapping -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
properties:{
{{- format_parameters(value['properties'], value['required'] | default([])) -}}
}
{%- elif value is mapping -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
properties:{
{{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
}
{%- endif -%}
{%- if value['required'] -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
required:[
{%- for item in value['required'] | default([]) -%}
<|"|>{{- item -}}<|"|>
{%- if not loop.last %},{% endif -%}
{%- endfor -%}
]
{%- endif -%}
{%- endif -%}
{%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
type:<|"|>{{ value['type'] | upper }}<|"|>}
{%- endif -%}
{%- endfor -%}
{%- endmacro -%}
{%- macro format_function_declaration(tool_data) -%}
declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
{%- set params = tool_data['function']['parameters'] -%}
{%- if params -%}
,parameters:{
{%- if params['properties'] -%}
properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
{%- endif -%}
{%- if params['required'] -%}
required:[
{%- for item in params['required'] -%}
<|"|>{{- item -}}<|"|>
{{- ',' if not loop.last -}}
{%- endfor -%}
],
{%- endif -%}
{%- if params['type'] -%}
type:<|"|>{{- params['type'] | upper -}}<|"|>}
{%- endif -%}
{%- endif -%}
{%- if 'response' in tool_data['function'] -%}
{%- set response_declaration = tool_data['function']['response'] -%}
,response:{
{%- if response_declaration['description'] -%}
description:<|"|>{{- response_declaration['description'] -}}<|"|>,
{%- endif -%}
{%- if response_declaration['type'] | upper == 'OBJECT' -%}
type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
{%- endif -%}
{%- endif -%}
}
{%- endmacro -%}
{%- macro format_argument(argument, escape_keys=True) -%}
{%- if argument is none -%}
{{- 'null' -}}
{%- elif argument is string -%}
{{- '<|"|>' + argument + '<|"|>' -}}
{%- elif argument is boolean -%}
{{- 'true' if argument else 'false' -}}
{%- elif argument is mapping -%}
{{- '{' -}}
{%- set ns = namespace(found_first=false) -%}
{%- for key, value in argument | dictsort -%}
{%- if ns.found_first %},{% endif -%}
{%- set ns.found_first = true -%}
{%- if escape_keys -%}
{{- '<|"|>' + key + '<|"|>' -}}
{%- else -%}
{{- key -}}
{%- endif -%}
:{{- format_argument(value, escape_keys=escape_keys) -}}
{%- endfor -%}
{{- '}' -}}
{%- elif argument is sequence -%}
{{- '[' -}}
{%- for item in argument -%}
{{- format_argument(item, escape_keys=escape_keys) -}}
{%- if not loop.last %},{% endif -%}
{%- endfor -%}
{{- ']' -}}
{%- else -%}
{{- argument -}}
{%- endif -%}
{%- endmacro -%}
{%- macro strip_thinking(text) -%}
{%- set ns = namespace(result='') -%}
{%- for part in text.split('<channel|>') -%}
{%- if '<|channel>' in part -%}
{%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
{%- else -%}
{%- set ns.result = ns.result + part -%}
{%- endif -%}
{%- endfor -%}
{{- ns.result | trim -}}
{%- endmacro -%}
{%- macro format_tool_response_block(tool_name, response) -%}
{{- '<|tool_response>' -}}
{%- if response is mapping -%}
{{- 'response:' + tool_name + '{' -}}
{%- for key, value in response | dictsort -%}
{{- key -}}:{{- format_argument(value, escape_keys=False) -}}
{%- if not loop.last %},{% endif -%}
{%- endfor -%}
{{- '}' -}}
{%- else -%}
{{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
{%- endif -%}
{{- '<tool_response|>' -}}
{%- endmacro -%}
{#- ===== SETUP ===== -#}
{%- set ns = namespace(prev_message_type=None, prev_non_tool_role=None) -%}
{%- set loop_messages = messages -%}
{%- set enable_thinking = enable_thinking | default(false) -%}
{%- set preserve_thinking = preserve_thinking | default(true) -%}
{{- bos_token -}}
{#- Handle System/Tool Definitions Block -#}
{%- if enable_thinking or tools or (messages and messages[0]['role'] in ['system', 'developer']) -%}
{{- '<|turn>system\n' -}}
{#- Inject Thinking token at the very top of the FIRST system turn -#}
{%- if enable_thinking -%}
{{- '<|think|>\n' -}}
{%- set ns.prev_message_type = 'think' -%}
{%- endif -%}
{%- if messages and messages[0]['role'] in ['system', 'developer'] -%}
{%- if messages[0]['content'] is string -%}
{{- messages[0]['content'] | trim -}}
{%- elif messages[0]['content'] is sequence -%}
{%- for item in messages[0]['content'] -%}
{{- item['text'] | trim + ' '-}}
{%- endfor -%}
{%- endif -%}
{%- set loop_messages = messages[1:] -%}
{%- endif -%}
{%- if tools -%}
{%- for tool in tools %}
{{- '<|tool>' -}}
{{- format_function_declaration(tool) | trim -}}
{{- '<tool|>' -}}
{%- endfor %}
{%- set ns.prev_message_type = 'tool' -%}
{%- endif -%}
{{- '<turn|>\n' -}}
{%- endif %}
{#- Pre-scan: find last user message index for reasoning guard -#}
{%- set ns_turn = namespace(last_user_idx=-1) -%}
{%- for i in range(loop_messages | length) -%}
{%- if loop_messages[i]['role'] == 'user' -%}
{%- set ns_turn.last_user_idx = i -%}
{%- endif -%}
{%- endfor -%}
{#- Loop through messages -#}
{%- for message in loop_messages -%}
{%- if message['role'] != 'tool' -%}
{%- set ns.prev_message_type = None -%}
{%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
{#- Detect continuation using tracked state — O(1) instead of O(n) backward scan -#}
{%- set continue_same_model_turn = (role == 'model' and ns.prev_non_tool_role == 'assistant') -%}
{%- if not continue_same_model_turn -%}
{{- '<|turn>' + role + '\n' }}
{%- endif -%}
{#- Render reasoning/reasoning_content as thinking channel (tool-call turns only) -#}
{%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
{%- set thinking_gate = (loop.index0 > ns_turn.last_user_idx) or preserve_thinking -%}
{%- if thinking_text and thinking_gate and message.get('tool_calls') -%}
{{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
{%- endif -%}
{%- if message.get('tool_calls') -%}
{%- for tool_call in message.get('tool_calls') -%}
{%- set function = tool_call['function'] -%}
{{- '<|tool_call>call:' + function['name'] + '{' -}}
{%- if function['arguments'] is mapping -%}
{%- set ns_args = namespace(found_first=false) -%}
{%- for key, value in function['arguments'] | dictsort -%}
{%- if ns_args.found_first %},{% endif -%}
{%- set ns_args.found_first = true -%}
{{- key -}}:{{- format_argument(value, escape_keys=False) -}}
{%- endfor -%}
{%- elif function['arguments'] is none -%}
{%- else -%}
{{- raise_exception(
"chat_template: tool_calls[].function.arguments must be a "
"JSON object (mapping), not a string. Deserialize arguments "
"before passing to the template."
) -}}
{%- endif -%}
{{- '}<tool_call|>' -}}
{%- endfor -%}
{%- set ns.prev_message_type = 'tool_call' -%}
{%- endif -%}
{%- set ns_tr_out = namespace(flag=false) -%}
{%- if message.get('tool_responses') -%}
{#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
{%- for tool_response in message.get('tool_responses') -%}
{{- format_tool_response_block(tool_response['name'] | default('unknown', true), tool_response['response']) -}}
{%- set ns_tr_out.flag = true -%}
{%- set ns.prev_message_type = 'tool_response' -%}
{%- endfor -%}
{%- elif message.get('tool_calls') -%}
{#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
{%- set ns_tool_scan = namespace(stopped=false) -%}
{%- for k in range(loop.index0 + 1, loop_messages | length) -%}
{%- if ns_tool_scan.stopped -%}
{%- elif loop_messages[k]['role'] != 'tool' -%}
{%- set ns_tool_scan.stopped = true -%}
{%- else -%}
{%- set follow = loop_messages[k] -%}
{#- Resolve tool_call_id to function name -#}
{%- set ns_tname = namespace(name=follow.get('name') or 'unknown') -%}
{%- for tc in message.get('tool_calls') -%}
{%- if tc.get('id') == follow.get('tool_call_id') -%}
{%- set ns_tname.name = tc['function']['name'] -%}
{%- endif -%}
{%- endfor -%}
{#- Handle content as string or content-parts array -#}
{%- set tool_body = follow.get('content') -%}
{%- if tool_body is string -%}
{{- format_tool_response_block(ns_tname.name, tool_body) -}}
{%- elif tool_body is sequence and tool_body is not string -%}
{%- set ns_txt = namespace(s='') -%}
{%- for part in tool_body -%}
{%- if part.get('type') == 'text' -%}
{%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
{%- endif -%}
{%- endfor -%}
{{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
{%- for part in tool_body -%}
{%- if part.get('type') in ['image', 'image_url'] -%}
{{- '<|image|>' -}}
{%- elif part.get('type') in ['audio', 'input_audio'] -%}
{{- '<|audio|>' -}}
{%- elif part.get('type') == 'video' -%}
{{- '<|video|>' -}}
{%- endif -%}
{%- endfor -%}
{%- else -%}
{{- format_tool_response_block(ns_tname.name, tool_body) -}}
{%- endif -%}
{%- set ns_tr_out.flag = true -%}
{%- set ns.prev_message_type = 'tool_response' -%}
{%- endif -%}
{%- endfor -%}
{%- endif -%}
{%- set captured_content -%}
{%- if message.get('content') is string -%}
{%- if role == 'model' -%}
{{- strip_thinking(message['content']) -}}
{%- else -%}
{{- message['content'] | trim -}}
{%- endif -%}
{%- elif message.get('content') is sequence -%}
{%- for item in message['content'] -%}
{%- if item.get('type') == 'text' -%}
{%- if role == 'model' -%}
{{- strip_thinking(item['text']) -}}
{%- else -%}
{{- item['text'] | trim -}}
{%- endif -%}
{%- elif item.get('type') in ['image', 'image_url'] -%}
{{- '<|image|>' -}}
{%- elif item.get('type') in ['audio', 'input_audio'] -%}
{{- '<|audio|>' -}}
{%- elif item.get('type') == 'video' -%}
{{- '<|video|>' -}}
{%- endif -%}
{%- endfor -%}
{%- endif -%}
{%- endset -%}
{{- captured_content -}}
{%- set has_content = captured_content | trim | length > 0 -%}
{#- Forward-scan: find next non-tool message role for continuation detection -#}
{%- set next_nt = namespace(role=None, found=false) -%}
{%- for j in range(loop.index0 + 1, loop_messages | length) -%}
{%- if not next_nt.found -%}
{%- if loop_messages[j]['role'] != 'tool' -%}
{%- set next_nt.role = loop_messages[j]['role'] -%}
{%- set next_nt.found = true -%}
{%- endif -%}
{%- endif -%}
{%- endfor -%}
{%- set continues_into_next = (
role == 'model'
and next_nt.role == 'assistant'
and not message.get('tool_calls')
and not ns_tr_out.flag
) -%}
{%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
{{- '<|tool_response>' -}}
{%- elif continues_into_next -%}
{#- {- '\n' -} -#}
{%- elif not (ns_tr_out.flag and not has_content) -%}
{{- '<turn|>\n' -}}
{%- endif -%}
{#- Track previous non-tool role for next iteration (avoids O(n) backward scan) -#}
{%- set ns.prev_non_tool_role = message['role'] -%}
{%- endif -%}
{%- endfor -%}
{%- if add_generation_prompt -%}
{%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
{{- '<|turn>model\n' -}}
{%- endif -%}
{%- if not enable_thinking -%}
{#- Suppress thinking — but not when awaiting tool responses -#}
{%- if ns.prev_message_type != 'tool_call' -%}
{{- '<|channel>thought\n<channel|>' -}}
{%- endif -%}
{%- endif -%}
{%- endif -%}
In my own experience I would say that small models (less than 20B active parameters) are prune to problems when using both large context and KV quantization, that increases with the higher quantization of the weights, of course. A good balance is around 64K cache and q8_0 for KV cache if VRAM limited, as in my case.
I hope it helps.
Edit: Only now I noticed you are a Google member, my bad. Two Martins in the same org is somewhat unexpected. Anyway, above is the result of my own empirical tests. I do like how Gemma4 handles writing and logic, but for coding and agentic tasks, besides frontend, it seems it still lacks a bit behind. Thank you a lot for providing Gemm4 weights with a permissive license. Regards.
I tried the fix from the above link and it did not solve the infinite thinking problem. The test was conducted in the Unsloth Studio.
I tried the fix from the above link and it did not solve the infinite thinking problem. The test was conducted in the Unsloth Studio.
the final chat template PR is this one: https://huggingface.co/google/gemma-4-26B-A4B-it/discussions/47
we tested multiple looping situations and they were all fixed with this PR.
can you please just try the final one and let me know if it fixes?
if not, it would be immensely helpful if you could share the prompt you are trying so we can see what is the issue (of course, only if doesn't have any private information)
I tested the latest chat template, but it still thinks indefinitely without reaching final answer.
Prompt: "Solve this differential equation: $[tg(x)*sin^2(y)]dx + [cos^2(x)*ctg(y)]dy = 0$"
Context Length: 8192
KV Cache Dtype: fp16
Speculative Decoding: off
Temperature: 1.0
Top P: 0.95
Top K: 64
Min P: 0.01
Repetition Penalty: off | 1.05
Presence Penalty: off
This comment has been hidden (marked as Spam)