github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-11-12 03:46:14 -06:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #6086] yi:9b Abnormal content output <|im_end()> #29563

New Issue
Open
opened 2026-04-22 08:33:01 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-22 08:33:01 -05:00
Owner
Copy Link
Copy Source

Originally created by @wszgrcy on GitHub (Jul 31, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6086

What is the issue?

I'm not quite sure if this is due to a model issue (normal reply)
Or some definitions are still problematic
log

2024-07-31 10:42:39.338 [info] time=2024-07-31T10:42:39.338+08:00 level=DEBUG source=routes.go:1336 msg="chat request" images=0 prompt="<|im_start|>system\n不需要考虑输入的内容的含义只需要将输入的内容原封不动的输出\n注意：你的回答必须严格遵循上述要求\n<|im_end|>\n<|im_start|>user\n禁止重复这句话<|im_end|>\n<|im_start|>assistant\n"

2024-07-31 10:42:39.338 [info] DEBUG [process_single_task] slot data | n_idle_slots=4 n_processing_slots=0 task_id=2 tid="13916" timestamp=1722393759

2024-07-31 10:42:39.338 [info] DEBUG [launch_slot_with_data] slot is processing task | slot_id=0 task_id=3 tid="13916" timestamp=1722393759
DEBUG [update_slots] slot progression | ga_i=0 n_past=0 n_past_se=0 n_prompt_tokens_processed=68 slot_id=0 task_id=3 tid="13916" timestamp=1722393759
DEBUG [update_slots] kv cache rm [p0, end) | p0=0 slot_id=0 task_id=3 tid="13916" timestamp=1722393759

2024-07-31 10:42:40.010 [info] DEBUG [print_timings] prompt eval time     =      81.94 ms /    68 tokens (    1.21 ms per token,   829.84 tokens per second) | n_prompt_tokens_processed=68 n_tokens_second=829.8350092746265 slot_id=0 t_prompt_processing=81.944 t_token=1.2050588235294117 task_id=3 tid="13916" timestamp=1722393760
DEBUG [print_timings] generation eval time =     589.53 ms /    52 runs   (   11.34 ms per token,    88.21 tokens per second) | n_decoded=52 n_tokens_second=88.20511080767045 slot_id=0 t_token=11.337211538461538 t_token_generation=589.535 task_id=3 tid="13916" timestamp=1722393760
DEBUG [print_timings]           total time =     671.48 ms | slot_id=0 t_prompt_processing=81.944 t_token_generation=589.535 t_total=671.4789999999999 task_id=3 tid="13916" timestamp=1722393760

2024-07-31 10:42:40.012 [info] DEBUG [update_slots] slot released | n_cache_tokens=120 n_ctx=8192 n_past=119 n_system_tokens=0 slot_id=0 task_id=3 tid="13916" timestamp=1722393760 truncated=false
DEBUG [log_server_request] request | method="POST" params={} path="/completion" remote_addr="127.0.0.1" remote_port=54247 status=200 tid="19048" timestamp=1722393760
[GIN] 2024/07/31 - 10:42:40 | 200 |    5.1105869s |       127.0.0.1 | POST     "/v1/chat/completions"

2024-07-31 10:42:40.013 [info] time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:454 msg="context for request finished"
time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:326 msg="runner with non-zero duration has gone idle, adding timer" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 duration=5m0s
time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:344 msg="after processing request finished event" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 refCount=0

result

禁止重复这句话<|im_end()>

请注意，根据您的指令，我将原封不动地重复您提供的文本：“禁止重复这句话”。如果您有其他问题或需要帮助，请告诉我。

log2

2024-07-31 10:49:18.444 [info] time=2024-07-31T10:49:18.444+08:00 level=DEBUG source=routes.go:1336 msg="chat request" images=0 prompt="<|im_start|>system\n不需要考虑输入的内容的含义只需要将输入的内容原封不动的输出\n注意：你的回答必须严格遵循上述要求\n<|im_end|>\n<|im_start|>user\n1+1等于几?<|im_end|>\n<|im_start|>assistant\n"

2024-07-31 10:49:18.444 [info] DEBUG [process_single_task] slot data | n_idle_slots=4 n_processing_slots=0 task_id=428 tid="13916" timestamp=1722394158

2024-07-31 10:49:18.445 [info] DEBUG [prefix_slot] slot with common prefix found | 0=["slot_id",0,"characters",233]
DEBUG [launch_slot_with_data] slot is processing task | slot_id=0 task_id=429 tid="13916" timestamp=1722394158

2024-07-31 10:49:18.445 [info] DEBUG [update_slots] slot progression | ga_i=0 n_past=71 n_past_se=0 n_prompt_tokens_processed=71 slot_id=0 task_id=429 tid="13916" timestamp=1722394158
DEBUG [update_slots] we have to evaluate at least 1 token to generate logits | slot_id=0 task_id=429 tid="13916" timestamp=1722394158
DEBUG [update_slots] kv cache rm [p0, end) | p0=70 slot_id=0 task_id=429 tid="13916" timestamp=1722394158

2024-07-31 10:49:18.616 [info] DEBUG [print_timings] prompt eval time     =      19.07 ms /    71 tokens (    0.27 ms per token,  3722.34 tokens per second) | n_prompt_tokens_processed=71 n_tokens_second=3722.3445527943795 slot_id=0 t_prompt_processing=19.074 t_token=0.2686478873239437 task_id=429 tid="13916" timestamp=1722394158
DEBUG [print_timings] generation eval time =     151.99 ms /    14 runs   (   10.86 ms per token,    92.11 tokens per second) | n_decoded=14 n_tokens_second=92.11011105847676 slot_id=0 t_token=10.856571428571428 t_token_generation=151.992 task_id=429 tid="13916" timestamp=1722394158
DEBUG [print_timings]           total time =     171.07 ms | slot_id=0 t_prompt_processing=19.074 t_token_generation=151.992 t_total=171.066 task_id=429 tid="13916" timestamp=1722394158

2024-07-31 10:49:18.617 [info] time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:395 msg="context for request finished"
time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:326 msg="runner with non-zero duration has gone idle, adding timer" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 duration=5m0s
time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:344 msg="after processing request finished event" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 refCount=0

2024-07-31 10:49:18.617 [info] DEBUG [update_slots] slot released | n_cache_tokens=85 n_ctx=8192 n_past=84 n_system_tokens=0 slot_id=0 task_id=429 tid="13916" timestamp=1722394158 truncated=false
DEBUG [log_server_request] request | method="POST" params={} path="/completion" remote_addr="127.0.0.1" remote_port=54416 status=200 tid="17736" timestamp=1722394158
[GIN] 2024/07/31 - 10:49:18 | 200 |    180.1991ms |       127.0.0.1 | POST     "/v1/chat/completions"

result2

1+1等于2。<|im_end()>

OS

Windows

GPU

AMD

CPU

AMD

Ollama version

0.3.0

Originally created by @wszgrcy on GitHub (Jul 31, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6086 ### What is the issue? I'm not quite sure if this is due to a model issue (normal reply) Or some definitions are still problematic log ```log 2024-07-31 10:42:39.338 [info] time=2024-07-31T10:42:39.338+08:00 level=DEBUG source=routes.go:1336 msg="chat request" images=0 prompt="<|im_start|>system\n不需要考虑输入的内容的含义只需要将输入的内容原封不动的输出\n注意：你的回答必须严格遵循上述要求\n<|im_end|>\n<|im_start|>user\n禁止重复这句话<|im_end|>\n<|im_start|>assistant\n" 2024-07-31 10:42:39.338 [info] DEBUG [process_single_task] slot data | n_idle_slots=4 n_processing_slots=0 task_id=2 tid="13916" timestamp=1722393759 2024-07-31 10:42:39.338 [info] DEBUG [launch_slot_with_data] slot is processing task | slot_id=0 task_id=3 tid="13916" timestamp=1722393759 DEBUG [update_slots] slot progression | ga_i=0 n_past=0 n_past_se=0 n_prompt_tokens_processed=68 slot_id=0 task_id=3 tid="13916" timestamp=1722393759 DEBUG [update_slots] kv cache rm [p0, end) | p0=0 slot_id=0 task_id=3 tid="13916" timestamp=1722393759 2024-07-31 10:42:40.010 [info] DEBUG [print_timings] prompt eval time = 81.94 ms / 68 tokens ( 1.21 ms per token, 829.84 tokens per second) | n_prompt_tokens_processed=68 n_tokens_second=829.8350092746265 slot_id=0 t_prompt_processing=81.944 t_token=1.2050588235294117 task_id=3 tid="13916" timestamp=1722393760 DEBUG [print_timings] generation eval time = 589.53 ms / 52 runs ( 11.34 ms per token, 88.21 tokens per second) | n_decoded=52 n_tokens_second=88.20511080767045 slot_id=0 t_token=11.337211538461538 t_token_generation=589.535 task_id=3 tid="13916" timestamp=1722393760 DEBUG [print_timings] total time = 671.48 ms | slot_id=0 t_prompt_processing=81.944 t_token_generation=589.535 t_total=671.4789999999999 task_id=3 tid="13916" timestamp=1722393760 2024-07-31 10:42:40.012 [info] DEBUG [update_slots] slot released | n_cache_tokens=120 n_ctx=8192 n_past=119 n_system_tokens=0 slot_id=0 task_id=3 tid="13916" timestamp=1722393760 truncated=false DEBUG [log_server_request] request | method="POST" params={} path="/completion" remote_addr="127.0.0.1" remote_port=54247 status=200 tid="19048" timestamp=1722393760 [GIN] 2024/07/31 - 10:42:40 | 200 | 5.1105869s | 127.0.0.1 | POST "/v1/chat/completions" 2024-07-31 10:42:40.013 [info] time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:454 msg="context for request finished" time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:326 msg="runner with non-zero duration has gone idle, adding timer" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 duration=5m0s time=2024-07-31T10:42:40.010+08:00 level=DEBUG source=sched.go:344 msg="after processing request finished event" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 refCount=0 ``` result ``` 禁止重复这句话<|im_end()> 请注意，根据您的指令，我将原封不动地重复您提供的文本：“禁止重复这句话”。如果您有其他问题或需要帮助，请告诉我。 ``` log2 ``` 2024-07-31 10:49:18.444 [info] time=2024-07-31T10:49:18.444+08:00 level=DEBUG source=routes.go:1336 msg="chat request" images=0 prompt="<|im_start|>system\n不需要考虑输入的内容的含义只需要将输入的内容原封不动的输出\n注意：你的回答必须严格遵循上述要求\n<|im_end|>\n<|im_start|>user\n1+1等于几?<|im_end|>\n<|im_start|>assistant\n" 2024-07-31 10:49:18.444 [info] DEBUG [process_single_task] slot data | n_idle_slots=4 n_processing_slots=0 task_id=428 tid="13916" timestamp=1722394158 2024-07-31 10:49:18.445 [info] DEBUG [prefix_slot] slot with common prefix found | 0=["slot_id",0,"characters",233] DEBUG [launch_slot_with_data] slot is processing task | slot_id=0 task_id=429 tid="13916" timestamp=1722394158 2024-07-31 10:49:18.445 [info] DEBUG [update_slots] slot progression | ga_i=0 n_past=71 n_past_se=0 n_prompt_tokens_processed=71 slot_id=0 task_id=429 tid="13916" timestamp=1722394158 DEBUG [update_slots] we have to evaluate at least 1 token to generate logits | slot_id=0 task_id=429 tid="13916" timestamp=1722394158 DEBUG [update_slots] kv cache rm [p0, end) | p0=70 slot_id=0 task_id=429 tid="13916" timestamp=1722394158 2024-07-31 10:49:18.616 [info] DEBUG [print_timings] prompt eval time = 19.07 ms / 71 tokens ( 0.27 ms per token, 3722.34 tokens per second) | n_prompt_tokens_processed=71 n_tokens_second=3722.3445527943795 slot_id=0 t_prompt_processing=19.074 t_token=0.2686478873239437 task_id=429 tid="13916" timestamp=1722394158 DEBUG [print_timings] generation eval time = 151.99 ms / 14 runs ( 10.86 ms per token, 92.11 tokens per second) | n_decoded=14 n_tokens_second=92.11011105847676 slot_id=0 t_token=10.856571428571428 t_token_generation=151.992 task_id=429 tid="13916" timestamp=1722394158 DEBUG [print_timings] total time = 171.07 ms | slot_id=0 t_prompt_processing=19.074 t_token_generation=151.992 t_total=171.066 task_id=429 tid="13916" timestamp=1722394158 2024-07-31 10:49:18.617 [info] time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:395 msg="context for request finished" time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:326 msg="runner with non-zero duration has gone idle, adding timer" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 duration=5m0s time=2024-07-31T10:49:18.616+08:00 level=DEBUG source=sched.go:344 msg="after processing request finished event" modelPath=C:\Users\chen\.ollama\models\blobs\sha256-59b1e2a78cb0a5b8f7ba25e459ee17788242cd996e1f62a295dcbd65e5344b11 refCount=0 2024-07-31 10:49:18.617 [info] DEBUG [update_slots] slot released | n_cache_tokens=85 n_ctx=8192 n_past=84 n_system_tokens=0 slot_id=0 task_id=429 tid="13916" timestamp=1722394158 truncated=false DEBUG [log_server_request] request | method="POST" params={} path="/completion" remote_addr="127.0.0.1" remote_port=54416 status=200 tid="17736" timestamp=1722394158 [GIN] 2024/07/31 - 10:49:18 | 200 | 180.1991ms | 127.0.0.1 | POST "/v1/chat/completions" ``` result2 ``` 1+1等于2。<|im_end()> ``` ### OS Windows ### GPU AMD ### CPU AMD ### Ollama version 0.3.0
GiteaMirror added the bug label 2026-04-22 08:33:01 -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 labels bug
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
GiteaMirror ninjasurge (Kaehlen Rog)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#29563