github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #1365] llama_print_timings have disappeared from the logs. #62753

New Issue
Closed
opened 2026-05-03 10:11:01 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-03 10:11:01 -05:00
Owner
Copy Link

Originally created by @madsamjp on GitHub (Dec 3, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1365

In a previous version of Ollama, following the logs (on Linux using journalctl -t ollama -f) would give helpful information after the model has finished with its response (such as tokens per second).

e.g. this:

Dec 03 14:58:42 osm-server ollama[20658]: llama server listening at http://127.0.0.1:54457
Dec 03 14:58:42 osm-server ollama[20658]: {"timestamp":1701615522,"level":"INFO","function":"main","line":1746,"message":"HTTP server listening","hostname":"127.0.0.1","port":54457}
Dec 03 14:58:42 osm-server ollama[20658]: {"timestamp":1701615522,"level":"INFO","function":"log_server_request","line":1233,"message":"request","remote_addr":"127.0.0.1","remote_port":51344,"statu>
Dec 03 14:58:42 osm-server ollama[937]: 2023/12/03 14:58:42 llama.go:492: llama runner started in 9.200880 seconds
Dec 03 14:58:50 osm-server ollama[20658]: {"timestamp":1701615530,"level":"INFO","function":"log_server_request","line":1233,"message":"request","remote_addr":"127.0.0.1","remote_port":51344,"statu>
Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings:        load time =    8317.76 ms
Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings:      sample time =     107.35 ms /   396 runs   (    0.27 ms per token,  3688.73 tokens per second)
Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: prompt eval time =     444.18 ms /   800 tokens (    0.56 ms per token,  1801.06 tokens per second)
Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings:        eval time =    6696.50 ms /   395 runs   (   16.95 ms per token,    58.99 tokens per second)
Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings:       total time =    7335.31 ms

This was really handy, but since updating Ollama, I've noticed this helpful info has gone. Is there an environment variable I can set to get it back?

Originally created by @madsamjp on GitHub (Dec 3, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1365 In a previous version of Ollama, following the logs (on Linux using `journalctl -t ollama -f`) would give helpful information after the model has finished with its response (such as tokens per second). e.g. this: ``` Dec 03 14:58:42 osm-server ollama[20658]: llama server listening at http://127.0.0.1:54457 Dec 03 14:58:42 osm-server ollama[20658]: {"timestamp":1701615522,"level":"INFO","function":"main","line":1746,"message":"HTTP server listening","hostname":"127.0.0.1","port":54457} Dec 03 14:58:42 osm-server ollama[20658]: {"timestamp":1701615522,"level":"INFO","function":"log_server_request","line":1233,"message":"request","remote_addr":"127.0.0.1","remote_port":51344,"statu> Dec 03 14:58:42 osm-server ollama[937]: 2023/12/03 14:58:42 llama.go:492: llama runner started in 9.200880 seconds Dec 03 14:58:50 osm-server ollama[20658]: {"timestamp":1701615530,"level":"INFO","function":"log_server_request","line":1233,"message":"request","remote_addr":"127.0.0.1","remote_port":51344,"statu> Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: load time = 8317.76 ms Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: sample time = 107.35 ms / 396 runs ( 0.27 ms per token, 3688.73 tokens per second) Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: prompt eval time = 444.18 ms / 800 tokens ( 0.56 ms per token, 1801.06 tokens per second) Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: eval time = 6696.50 ms / 395 runs ( 16.95 ms per token, 58.99 tokens per second) Dec 03 14:58:50 osm-server ollama[937]: llama_print_timings: total time = 7335.31 ms ``` This was really handy, but since updating Ollama, I've noticed this helpful info has gone. Is there an environment variable I can set to get it back?
GiteaMirror commented 2026-05-03 10:11:05 -05:00
Author
Owner
Copy Link

@rankun203 commented on GitHub (Dec 5, 2023):

As a workaround, you can use ollama run --verbose <MODEL_NAME> ...

Detailed usage of ollama run:

$ ollama run --help
Run a model

Usage:
  ollama run MODEL [PROMPT] [flags]

Flags:
      --format string   Response format (e.g. json)
  -h, --help            help for run
      --insecure        Use an insecure registry
      --nowordwrap      Don't wrap words to the next line automatically
      --verbose         Show timings for response

Example response:

... LLM output

total duration:       5.33640075s
load duration:        531.536292ms
prompt eval count:    27 token(s)
prompt eval duration: 110.818ms
prompt eval rate:     243.64 tokens/s
eval count:           270 token(s)
eval duration:        4.689066s
eval rate:            57.58 tokens/s
<!-- gh-comment-id:1840729645 --> @rankun203 commented on GitHub (Dec 5, 2023): As a workaround, you can use `ollama run --verbose <MODEL_NAME> ...` Detailed usage of `ollama run`: ```bash $ ollama run --help Run a model Usage: ollama run MODEL [PROMPT] [flags] Flags: --format string Response format (e.g. json) -h, --help help for run --insecure Use an insecure registry --nowordwrap Don't wrap words to the next line automatically --verbose Show timings for response ``` Example response: ```bash ... LLM output total duration: 5.33640075s load duration: 531.536292ms prompt eval count: 27 token(s) prompt eval duration: 110.818ms prompt eval rate: 243.64 tokens/s eval count: 270 token(s) eval duration: 4.689066s eval rate: 57.58 tokens/s ```
GiteaMirror commented 2026-05-03 10:11:06 -05:00
Author
Owner
Copy Link

@madsamjp commented on GitHub (Dec 9, 2023):

As a workaround, you can use ollama run --verbose <MODEL_NAME> ...

Detailed usage of ollama run:

$ ollama run --help
Run a model

Usage:
  ollama run MODEL [PROMPT] [flags]

Flags:
      --format string   Response format (e.g. json)
  -h, --help            help for run
      --insecure        Use an insecure registry
      --nowordwrap      Don't wrap words to the next line automatically
      --verbose         Show timings for response

Example response:

... LLM output

total duration:       5.33640075s
load duration:        531.536292ms
prompt eval count:    27 token(s)
prompt eval duration: 110.818ms
prompt eval rate:     243.64 tokens/s
eval count:           270 token(s)
eval duration:        4.689066s
eval rate:            57.58 tokens/s

This would only be useful when running a model in the terminal. If I run the models externally through the api, is there a way to get response timings?

<!-- gh-comment-id:1848636064 --> @madsamjp commented on GitHub (Dec 9, 2023): > As a workaround, you can use `ollama run --verbose <MODEL_NAME> ...` > > Detailed usage of `ollama run`: > > ```shell > $ ollama run --help > Run a model > > Usage: > ollama run MODEL [PROMPT] [flags] > > Flags: > --format string Response format (e.g. json) > -h, --help help for run > --insecure Use an insecure registry > --nowordwrap Don't wrap words to the next line automatically > --verbose Show timings for response > ``` > > Example response: > > ```shell > ... LLM output > > total duration: 5.33640075s > load duration: 531.536292ms > prompt eval count: 27 token(s) > prompt eval duration: 110.818ms > prompt eval rate: 243.64 tokens/s > eval count: 270 token(s) > eval duration: 4.689066s > eval rate: 57.58 tokens/s > ``` This would only be useful when running a model in the terminal. If I run the models externally through the api, is there a way to get response timings?
GiteaMirror commented 2026-05-03 10:11:07 -05:00
Author
Owner
Copy Link

@mxyng commented on GitHub (Jan 20, 2024):

the terminal metrics are based on generate or chat responses. responses will contain these fields:

total_duration: time spent generating the response
load_duration: time spent in nanoseconds loading the model
prompt_eval_count: number of tokens in the prompt
prompt_eval_duration: time spent in nanoseconds evaluating the prompt
eval_count: number of tokens the response
eval_duration: time in nanoseconds spent generating the response
<!-- gh-comment-id:1901398968 --> @mxyng commented on GitHub (Jan 20, 2024): the terminal metrics are based on [generate](https://github.com/jmorganca/ollama/blob/main/docs/api.md#generate-a-completion) or [chat](https://github.com/jmorganca/ollama/blob/main/docs/api.md#generate-a-chat-completion) responses. responses will contain these fields: ``` total_duration: time spent generating the response load_duration: time spent in nanoseconds loading the model prompt_eval_count: number of tokens in the prompt prompt_eval_duration: time spent in nanoseconds evaluating the prompt eval_count: number of tokens the response eval_duration: time in nanoseconds spent generating the response ```
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
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#62753
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?