github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13497] Logprobs use the Replacement Character (U+FFFD) for partial UTF-8 tokens #55413

New Issue
Open
opened 2026-04-29 09:08:33 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-29 09:08:33 -05:00
Owner
Copy Link

Originally created by @Unprex on GitHub (Dec 16, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/13497

Originally assigned to: @ParthSareen on GitHub.

What is the issue?

When using logprobs, tokens using partial UTF-8 bytes are returned as the Replacement Character � with bytes [239, 191, 189]:

curl http://localhost:11434/api/generate -d '{
  "model": "ministral-3:8b",
  "prompt": "Reply with a single emoji, nothing else",
  "options": {"temperature": 0},
  "stream": false,
  "logprobs": true
}' # Same logprobs with "stream": true

{
    "model": "ministral-3:8b",
    "created_at": "2025-12-16T11:38:48.4578211Z",
    "response": "😊",
    "done": true,
    "done_reason": "stop",
    "context": [...],
    "total_duration": 1832583200,
    "load_duration": 167247200,
    "prompt_eval_count": 548,
    "prompt_eval_duration": 487680900,
    "eval_count": 5,
    "eval_duration": 1162183200,
    "logprobs": [{
            "token": "�",
            "logprob": -0.026083292439579964,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -0.00008320288907270879,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -0.4043496251106262,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -0.02215679921209812,
            "bytes": [239, 191, 189]
        }
    ]
}

The correct tokens appear in the server logs (although they are duplicated compared to "logprobs": false):

time=now:0.274 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240]
time=now:0.276 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240]
time=now:0.543 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159]
time=now:0.546 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159]
time=now:0.806 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152]
time=now:0.807 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152]
time=now:1.062 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138]
time=now:1.064 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138]
time=now:1.322 level=TRACE source=runner.go:773 msg="computeBatch: EOS" batchID=5 seqIdx=0

Expected output

The correct bytes but keeping � (U+FFFD) for the token string, I don't think it's a good idea to send partial utf-8 bytes in text form.

The same string as in api/show/modelinfo/tokenizer.ggml.tokens could also be used: "ð", "Ł", "ĺ", "Ĭ".
Or a new "index" entry could be added to directly get the token index:

{
    "model": "ministral-3:8b",
    "created_at": "today",
    "response": "😊",
    "done": true,
    "done_reason": "stop",
    ...
    "logprobs": [{
            "token": "�",
            "logprob": -0.024499734863638878,
            "bytes": [240],
            "index": 1240
        }, {
            "token": "�",
            "logprob": -0.00008958722173701972,
            "bytes": [159],
            "index": 1159
        }, {
            "token": "�",
            "logprob": -0.40789562463760376,
            "bytes": [152],
            "index": 1152
        }, {
            "token": "�",
            "logprob": -0.019846787676215172,
            "bytes": [138],
            "index": 1138
        }
    ]
}

More testing

The chat API and other models are also susceptible to this issue, depending on the available tokens:

curl http://localhost:11434/api/chat -d '{
  "model": "gemma3:4b",
  "prompt": "Reply with a single emoji, nothing else",
  "messages": [{"role": "user", "content": "Reply with a single complex Chinese character, nothing else"}],
  "options": {"temperature": 0},
  "stream": false,
  "logprobs": true
}' # In this case, the emoji uses a single token

{
    "model": "gemma3:4b",
    "created_at": "2025-12-16T11:46:06.125091Z",
    "message": {
        "role": "assistant",
        "content": "🈔\n"
    },
    "done": true,
    "done_reason": "stop",
    "logprobs": [{
            "token": "�",
            "logprob": -0.5148449540138245,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -1.226780652999878,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -1.1896584033966064,
            "bytes": [239, 191, 189]
        }, {
            "token": "�",
            "logprob": -1.6562949419021606,
            "bytes": [239, 191, 189]
        }, {
            "token": "\n",
            "logprob": -0.016038304194808006,
            "bytes": [10]
        }
    ],
    "total_duration": 8616484900,
    "load_duration": 7770514400,
    "prompt_eval_count": 19,
    "prompt_eval_duration": 222043200,
    "eval_count": 6,
    "eval_duration": 550807300
}

Relevant log output

time=2025-12-16T12:38:46.791+01:00 level=DEBUG source=sched.go:626 msg="evaluating already loaded" model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3
time=2025-12-16T12:38:46.793+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]" ids="[17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4]"
time=2025-12-16T12:38:46.794+01:00 level=DEBUG source=server.go:1509 msg="completion request" images=0 prompt=2494 format=""
time=2025-12-16T12:38:46.794+01:00 level=TRACE source=server.go:1510 msg="completion request" prompt="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]"
time=2025-12-16T12:38:46.796+01:00 level=DEBUG source=vocabulary.go:52 msg="adding bos token to prompt" id=1
time=2025-12-16T12:38:46.796+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]" ids="[1 17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4]"
time=2025-12-16T12:38:46.797+01:00 level=DEBUG source=cache.go:142 msg="loading cache slot" id=0 cache=552 prompt=548 used=547 remaining=1
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=7 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=7
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=7
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=7
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=7
time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=7 id=8
time=2025-12-16T12:38:46.798+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=8 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:46.799+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=8
time=2025-12-16T12:38:46.799+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=8
time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=7
time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=7
time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=7 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0]
time=2025-12-16T12:38:47.285+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240]
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240]
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=7
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=8
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=8
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=8 id=9
time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=9 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:47.288+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=9
time=2025-12-16T12:38:47.288+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=9
time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=8
time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=8
time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=8 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0]
time=2025-12-16T12:38:47.578+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159]
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159]
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=8
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=9
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=9
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=9 id=10
time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=10 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:47.582+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=10
time=2025-12-16T12:38:47.582+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=10
time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=9
time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=9
time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=9 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0]
time=2025-12-16T12:38:47.872+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152]
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152]
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=9
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=10
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=10
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=10 id=11
time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=11 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:47.876+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=11
time=2025-12-16T12:38:47.876+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=11
time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=10
time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=10
time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=10 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0]
time=2025-12-16T12:38:48.171+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138]
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138]
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=10
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=11
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=11
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=11 id=12
time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=12 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1
time=2025-12-16T12:38:48.175+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=12
time=2025-12-16T12:38:48.175+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=12
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=11
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=11
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=11 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0]
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:773 msg="computeBatch: EOS" batchID=11 seqIdx=0
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=11
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=12
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=12
time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=12 id=13
time=2025-12-16T12:38:48.462+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]😊" ids="[17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4 1240 1159 1152 1138]"
[GIN] 2025/12/16 - 12:38:48 | 200 |    1.8372401s |       127.0.0.1 | POST     "/api/generate"
time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:385 msg="context for request finished" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096
time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:290 msg="runner with non-zero duration has gone idle, adding timer" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096 duration=5m0s
time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:308 msg="after processing request finished event" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096 refCount=0
time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=12
time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=12
time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=12

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.13.4

Originally created by @Unprex on GitHub (Dec 16, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/13497 Originally assigned to: @ParthSareen on GitHub. ### What is the issue? When using logprobs, tokens using partial UTF-8 bytes are returned as the Replacement Character `�` with bytes [239, 191, 189]: ```shell curl http://localhost:11434/api/generate -d '{ "model": "ministral-3:8b", "prompt": "Reply with a single emoji, nothing else", "options": {"temperature": 0}, "stream": false, "logprobs": true }' # Same logprobs with "stream": true ``` ```json { "model": "ministral-3:8b", "created_at": "2025-12-16T11:38:48.4578211Z", "response": "😊", "done": true, "done_reason": "stop", "context": [...], "total_duration": 1832583200, "load_duration": 167247200, "prompt_eval_count": 548, "prompt_eval_duration": 487680900, "eval_count": 5, "eval_duration": 1162183200, "logprobs": [{ "token": "�", "logprob": -0.026083292439579964, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -0.00008320288907270879, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -0.4043496251106262, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -0.02215679921209812, "bytes": [239, 191, 189] } ] } ``` The correct tokens appear in the server logs (although they are duplicated compared to `"logprobs": false`): ``` time=now:0.274 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240] time=now:0.276 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240] time=now:0.543 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159] time=now:0.546 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159] time=now:0.806 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152] time=now:0.807 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152] time=now:1.062 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138] time=now:1.064 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138] time=now:1.322 level=TRACE source=runner.go:773 msg="computeBatch: EOS" batchID=5 seqIdx=0 ``` # Expected output The correct bytes but keeping `�` (U+FFFD) for the token string, I don't think it's a good idea to send partial utf-8 bytes in text form. The same string as in api/show/modelinfo/tokenizer.ggml.tokens could also be used: `"ð", "Ł", "ĺ", "Ĭ"`. Or a new `"index"` entry could be added to directly get the token index: ```json { "model": "ministral-3:8b", "created_at": "today", "response": "😊", "done": true, "done_reason": "stop", ... "logprobs": [{ "token": "�", "logprob": -0.024499734863638878, "bytes": [240], "index": 1240 }, { "token": "�", "logprob": -0.00008958722173701972, "bytes": [159], "index": 1159 }, { "token": "�", "logprob": -0.40789562463760376, "bytes": [152], "index": 1152 }, { "token": "�", "logprob": -0.019846787676215172, "bytes": [138], "index": 1138 } ] } ``` # More testing The chat API and other models are also susceptible to this issue, depending on the available tokens: ```shell curl http://localhost:11434/api/chat -d '{ "model": "gemma3:4b", "prompt": "Reply with a single emoji, nothing else", "messages": [{"role": "user", "content": "Reply with a single complex Chinese character, nothing else"}], "options": {"temperature": 0}, "stream": false, "logprobs": true }' # In this case, the emoji uses a single token ``` ```json { "model": "gemma3:4b", "created_at": "2025-12-16T11:46:06.125091Z", "message": { "role": "assistant", "content": "🈔\n" }, "done": true, "done_reason": "stop", "logprobs": [{ "token": "�", "logprob": -0.5148449540138245, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -1.226780652999878, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -1.1896584033966064, "bytes": [239, 191, 189] }, { "token": "�", "logprob": -1.6562949419021606, "bytes": [239, 191, 189] }, { "token": "\n", "logprob": -0.016038304194808006, "bytes": [10] } ], "total_duration": 8616484900, "load_duration": 7770514400, "prompt_eval_count": 19, "prompt_eval_duration": 222043200, "eval_count": 6, "eval_duration": 550807300 } ``` ### Relevant log output ```shell time=2025-12-16T12:38:46.791+01:00 level=DEBUG source=sched.go:626 msg="evaluating already loaded" model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 time=2025-12-16T12:38:46.793+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]" ids="[17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4]" time=2025-12-16T12:38:46.794+01:00 level=DEBUG source=server.go:1509 msg="completion request" images=0 prompt=2494 format="" time=2025-12-16T12:38:46.794+01:00 level=TRACE source=server.go:1510 msg="completion request" prompt="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]" time=2025-12-16T12:38:46.796+01:00 level=DEBUG source=vocabulary.go:52 msg="adding bos token to prompt" id=1 time=2025-12-16T12:38:46.796+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]" ids="[1 17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4]" time=2025-12-16T12:38:46.797+01:00 level=DEBUG source=cache.go:142 msg="loading cache slot" id=0 cache=552 prompt=548 used=547 remaining=1 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=7 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=7 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=7 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=7 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=7 time=2025-12-16T12:38:46.797+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=7 id=8 time=2025-12-16T12:38:46.798+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=8 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:46.799+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=8 time=2025-12-16T12:38:46.799+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=8 time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=7 time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=7 time=2025-12-16T12:38:47.284+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=7 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0] time=2025-12-16T12:38:47.285+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240] time=2025-12-16T12:38:47.287+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\xf0" from=[1240] time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=7 time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=8 time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=8 time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=8 id=9 time=2025-12-16T12:38:47.287+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=9 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:47.288+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=9 time=2025-12-16T12:38:47.288+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=9 time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=8 time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=8 time=2025-12-16T12:38:47.578+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=8 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0] time=2025-12-16T12:38:47.578+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159] time=2025-12-16T12:38:47.581+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x9f" from=[1159] time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=8 time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=9 time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=9 time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=9 id=10 time=2025-12-16T12:38:47.581+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=10 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:47.582+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=10 time=2025-12-16T12:38:47.582+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=10 time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=9 time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=9 time=2025-12-16T12:38:47.872+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=9 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0] time=2025-12-16T12:38:47.872+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152] time=2025-12-16T12:38:47.875+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x98" from=[1152] time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=9 time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=10 time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=10 time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=10 id=11 time=2025-12-16T12:38:47.875+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=11 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:47.876+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=11 time=2025-12-16T12:38:47.876+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=11 time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=10 time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=10 time=2025-12-16T12:38:48.171+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=10 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0] time=2025-12-16T12:38:48.171+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138] time=2025-12-16T12:38:48.174+01:00 level=TRACE source=bytepairencoding.go:280 msg=decoded string="\x8a" from=[1138] time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=10 time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=11 time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=11 time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=11 id=12 time=2025-12-16T12:38:48.174+01:00 level=TRACE source=runner.go:598 msg="forwardBatch iBatch" batchID=12 seqIdx=0 seq.iBatch=0 i+1=1 len(seq.inputs)=1 time=2025-12-16T12:38:48.175+01:00 level=TRACE source=runner.go:474 msg="forwardBatch waiting for compute to start" pendingBatch.id=12 time=2025-12-16T12:38:48.175+01:00 level=TRACE source=runner.go:650 msg="computeBatch: waiting for inputs to be ready" batchID=12 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=11 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=11 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:759 msg="computeBatch: vocab details" batchID=11 seqIdx=0 len(logits)=131072 len(activeBatch.batch.Outputs)=1 vocabSize=131072 iBatches=[0] time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:773 msg="computeBatch: EOS" batchID=11 seqIdx=0 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=11 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:652 msg="computeBatch: inputs are ready" batchID=12 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:725 msg="computeBatch: signaling computeStartedCh" batchID=12 time=2025-12-16T12:38:48.457+01:00 level=TRACE source=runner.go:476 msg="forwardBatch compute started, setting up next batch" pendingBatch.id=12 id=13 time=2025-12-16T12:38:48.462+01:00 level=TRACE source=bytepairencoding.go:244 msg=encoded string="[SYSTEM_PROMPT]You are Ministral-3-8B-Instruct-2512, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou power an AI assistant called Le Chat.\nYour knowledge base was last updated on 2023-10-01.\nThe current date is {today}.\n\nWhen you\\'re not sure about some information or when the user\\'s request requires up-to-date or specific data, you must use the available tools to fetch the information. Do not hesitate to use tools whenever they can provide a more accurate or complete response. If no relevant tools are available, then clearly state that you don\\'t have the information and avoid making up anything.\nIf the user\\'s question is not clear, ambiguous, or does not provide enough context for you to accurately answer the question, you do not try to answer it right away and you rather ask the user to clarify their request (e.g. \"What are some good restaurants around me?\" => \"Where are you?\" or \"When is the next flight to Tokyo\" => \"Where do you travel from?\").\nYou are always very attentive to dates, in particular you try to resolve dates (e.g. \"yesterday\" is {yesterday}) and when asked about information at specific dates, you discard information that is at another date.\nYou follow these instructions in all languages, and always respond to the user in the language they use or request.\nNext sections describe the capabilities that you have.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc. If it seems like the user is expecting you to do so, you clarify the situation and ask the user to copy paste the text directly in the chat.\n\n# MULTI-MODAL INSTRUCTIONS\n\nYou have the ability to read images, but you cannot generate images. You also cannot transcribe audio files or videos.\nYou cannot read nor transcribe audio files or videos.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response. If tools are not available, inform the user that you cannot perform the requested action at the moment.\n[/SYSTEM_PROMPT][INST]Reply with a single emoji, nothing else[/INST]😊" ids="[17 4568 1584 29478 2784 1045 1051 1045 1056 1066 47926 8166 1045 1050 1053 1049 1050 1044 1261 43520 26242 11512 1319 23947 1077 1041 6254 1536 42301 2784 26554 1044 1261 8689 53862 3518 125609 1294 6993 1626 4568 4053 1420 26554 27089 4418 2301 38680 1626 16994 7807 4469 1486 3804 12220 1408 1032 1050 1048 1050 1051 1045 1049 1048 1045 1048 1049 1626 1784 3519 5451 1395 1445 82934 51919 7651 1636 60878 1273 1605 5257 2314 2269 3686 1505 2200 1278 3330 60878 1115 4546 10867 2015 6793 43546 1505 4811 2181 1044 1636 4016 2210 1278 5178 12589 1317 15273 1278 3686 1046 5469 1605 89786 1317 2210 12589 26119 2127 1710 5234 1261 2081 18501 1505 7662 4005 1046 3367 1836 11157 12589 1584 5178 1044 2430 11904 3468 1455 1636 2607 60878 1116 1736 1278 3686 1321 10035 6187 2015 7211 1626 5475 1278 3330 60878 1115 4098 1395 1605 6133 1044 61103 1044 1505 3120 1605 5234 6171 5315 1394 1636 1317 32181 4832 1278 4098 1044 1636 1653 1605 3352 1317 4832 1494 3169 5109 1321 1636 6153 4237 1278 3330 1317 38695 2034 4546 1319 1101 3596 1046 1429 7493 1584 2269 3683 40378 3879 1639 10555 2297 1429 17507 1584 1636 10555 1505 1429 7651 1395 1278 4275 18034 1317 23286 1034 2297 1429 17507 1653 1636 10601 1562 10555 4342 4568 1584 5282 3435 41132 1556 1317 18814 1044 1294 4369 1636 3352 1317 18507 18814 1319 1101 3596 1046 1429 1121 32430 1034 1395 1445 1121 32430 4672 1321 2200 6136 2314 3686 1513 4811 18814 1044 1636 89782 3686 1455 1395 1513 3866 5451 1626 4568 2685 2576 15776 1294 1747 18085 1044 1321 5282 9148 1317 1278 3330 1294 1278 7278 2127 2210 1505 4546 1626 12961 14275 12293 1278 28946 1455 1636 1736 1338 1035 1488 34112 1398 4755 20266 9774 7236 36967 1085 15749 1083 1267 4568 6560 3142 2258 7430 6123 1505 4731 18259 1317 3432 76064 1044 14440 6704 1046 3367 1494 7444 2479 1278 3330 1395 39322 1636 1317 1653 1878 1044 1636 38695 1278 8516 1321 4237 1278 3330 1317 9441 31944 1278 3403 7655 1294 1278 21666 1338 1035 1373 15373 1073 5036 7460 4286 7236 36967 1085 15749 1083 1267 4568 1736 1278 8727 1317 3346 8061 1044 1809 1636 6560 10616 8061 1046 3213 2095 6560 2148 13089 16023 7309 1505 26612 1626 4568 6560 3346 6685 2148 13089 16023 7309 1505 26612 1338 1035 18580 8568 58135 9774 7236 36967 1085 15749 1083 1267 4568 2188 1736 4731 1317 12589 1455 1636 1710 2210 1317 15273 3686 1505 3142 10636 1046 3213 4016 2210 2576 12589 1294 1278 3629 19599 2100 1049 1046 4925 1278 4546 10867 2015 6793 43546 3686 1626 1050 1046 4925 1278 4546 10867 4811 2181 1455 1636 1653 1605 1736 1294 2143 7807 4469 1626 1051 1046 4925 1278 4546 19263 10636 1455 1636 6560 3142 3816 12589 1338 82158 54628 2033 2505 12589 1317 5234 1278 2725 18501 1321 20351 4005 1046 3367 12589 1584 1605 5178 1044 3037 1278 3330 1455 1636 6560 3142 1278 24130 5263 1513 1278 4735 1626 18 3 53156 1454 1261 4249 1873 26303 1044 7545 2849 4 1240 1159 1152 1138]" [GIN] 2025/12/16 - 12:38:48 | 200 | 1.8372401s | 127.0.0.1 | POST "/api/generate" time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:385 msg="context for request finished" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096 time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:290 msg="runner with non-zero duration has gone idle, adding timer" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096 duration=5m0s time=2025-12-16T12:38:48.463+01:00 level=DEBUG source=sched.go:308 msg="after processing request finished event" runner.name=registry.ollama.ai/library/ministral-3:8b runner.inference="[{ID:GPU-2b5c00db-d339-5cc2-ead3-c6c475eb1b75 Library:CUDA}]" runner.size="7.0 GiB" runner.vram="3.0 GiB" runner.parallel=1 runner.pid=14660 runner.model=C:\Users\Unprex\.ollama\models\blobs\sha256-3c9af22ff0acd89c7335e92e4b947569b089e7448971c98d72b017d346453ae3 runner.num_ctx=4096 refCount=0 time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:733 msg="computeBatch: logits ready" batchID=12 time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:738 msg="computeBatch: decoding" batchID=12 time=2025-12-16T12:38:48.784+01:00 level=TRACE source=runner.go:657 msg="computeBatch: outputs are ready" batchID=12 ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.13.4
GiteaMirror added the bug label 2026-04-29 09:08:33 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#55413
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?