github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-05 23:53:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2732] Provide model context length in API #1642

New Issue
Closed
opened 2026-04-12 11:35:21 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 11:35:21 -05:00
Owner
Copy Link

Originally created by @gluonfield on GitHub (Feb 24, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2732

Hi guys, love Ollama and contributing to the ecosystem by building Enchanted.

One important thing that is currently missing from /api/show API is the context length that model supports. For all RAG applications this is essential to know and it seems that in the future models will support greatly varied context lengths.

It would be great to include this metadata into API and as well as Ollama library.

Example of /api/show response for Gemma contains no clues of context length.

{
  "license": "...",
  "modelfile": "# Modelfile generated by \"ollama show\"\n# To build a new Modelfile based on this one, replace the FROM line with:\n# FROM gemma:7b\n\nFROM /Users/wpc/.ollama/models/blobs/sha256:456402914e838a953e0cf80caa6adbe75383d9e63584a964f504a7bbb8f7aad9\nTEMPLATE \"\"\"<start_of_turn>user\n{{ if .System }}{{ .System }} {{ end }}{{ .Prompt }}<end_of_turn>\n<start_of_turn>model\n{{ .Response }}<end_of_turn>\n\"\"\"\nPARAMETER repeat_penalty 1\nPARAMETER stop \"<start_of_turn>\"\nPARAMETER stop \"<end_of_turn>\"",
  "parameters": "repeat_penalty                 1\nstop                           \"<start_of_turn>\"\nstop                           \"<end_of_turn>\"",
  "template": "<start_of_turn>user\n{{ if .System }}{{ .System }} {{ end }}{{ .Prompt }}<end_of_turn>\n<start_of_turn>model\n{{ .Response }}<end_of_turn>\n",
  "details": {
    "parent_model": "",
    "format": "gguf",
    "family": "gemma",
    "families": [
      "gemma"
    ],
    "parameter_size": "9B",
    "quantization_level": "Q4_0"
  }
}
Originally created by @gluonfield on GitHub (Feb 24, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2732 Hi guys, love Ollama and contributing to the ecosystem by building [Enchanted](https://github.com/AugustDev/enchanted). One important thing that is currently missing from `/api/show` API is the context length that model supports. For all RAG applications this is essential to know and it seems that in the future models will support greatly varied context lengths. It would be great to include this metadata into API and as well as [Ollama library](https://ollama.com/library). Example of `/api/show` response for Gemma contains no clues of context length. ``` { "license": "...", "modelfile": "# Modelfile generated by \"ollama show\"\n# To build a new Modelfile based on this one, replace the FROM line with:\n# FROM gemma:7b\n\nFROM /Users/wpc/.ollama/models/blobs/sha256:456402914e838a953e0cf80caa6adbe75383d9e63584a964f504a7bbb8f7aad9\nTEMPLATE \"\"\"<start_of_turn>user\n{{ if .System }}{{ .System }} {{ end }}{{ .Prompt }}<end_of_turn>\n<start_of_turn>model\n{{ .Response }}<end_of_turn>\n\"\"\"\nPARAMETER repeat_penalty 1\nPARAMETER stop \"<start_of_turn>\"\nPARAMETER stop \"<end_of_turn>\"", "parameters": "repeat_penalty 1\nstop \"<start_of_turn>\"\nstop \"<end_of_turn>\"", "template": "<start_of_turn>user\n{{ if .System }}{{ .System }} {{ end }}{{ .Prompt }}<end_of_turn>\n<start_of_turn>model\n{{ .Response }}<end_of_turn>\n", "details": { "parent_model": "", "format": "gguf", "family": "gemma", "families": [ "gemma" ], "parameter_size": "9B", "quantization_level": "Q4_0" } } ```
GiteaMirror added the feature request label 2026-04-12 11:35:22 -05:00
GiteaMirror commented 2026-04-12 11:35:24 -05:00
Author
Owner
Copy Link

@Nurgo commented on GitHub (Mar 20, 2024):

I can confirm that this feature would be very useful. We'd love to add Ollama support to BrainSoup, but not being able to know the maximum context size of models is very problematic for optimizing RAG.

<!-- gh-comment-id:2009281697 --> @Nurgo commented on GitHub (Mar 20, 2024): I can confirm that this feature would be very useful. We'd love to add Ollama support to [BrainSoup](https://www.nurgo-software.com/products/brainsoup), but not being able to know the maximum context size of models is very problematic for optimizing RAG.
GiteaMirror commented 2026-04-12 11:35:26 -05:00
Author
Owner
Copy Link

@supercurio commented on GitHub (Jun 4, 2024):

I need this feature as well, in order to set the context size and build the chat history according to the model's capability.

Current workarounds have shortcomings:

  • pattern matching based on the name is not ideal when llama3 is 8k and llama3-gradient is 1048k. But feasible when matching for "llama3:" or "llama3-gradient:" strings.
  • pattern matching on the base name doesn't work for phi3 which exists in 4k and 128k variants.
  • the info is available on ollama.com but it doesn't work offline.

The maximum context size info itself would be valuable. It might be even more useful in conjunction in an API which returns also:

  • VRAM availability
  • Which model + context size fits in VRAM.
<!-- gh-comment-id:2147640966 --> @supercurio commented on GitHub (Jun 4, 2024): I need this feature as well, in order to set the context size and build the chat history according to the model's capability. Current workarounds have shortcomings: - pattern matching based on the name is not ideal when llama3 is 8k and llama3-gradient is 1048k. But feasible when matching for "llama3:" or "llama3-gradient:" strings. - pattern matching on the base name doesn't work for phi3 which exists in 4k and 128k variants. - the info is [available on ollama.com](https://ollama.com/library/llama3-gradient/blobs/011c3962dbd7) but it doesn't work offline. The maximum context size info itself would be valuable. It might be even more useful in conjunction in an API which returns also: - VRAM availability - Which model + context size fits in VRAM.
GiteaMirror commented 2026-04-12 11:35:26 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Jun 25, 2024):

Hi all, this should be possible via /api/show:

"model_info": {
    "general.architecture": "llama",
    "general.file_type": 2,
    "general.parameter_count": 8030261248,
    "general.quantization_version": 2,
    "llama.attention.head_count": 32,
    "llama.attention.head_count_kv": 8,
    "llama.attention.layer_norm_rms_epsilon": 0.00001,
    "llama.block_count": 32,
    "llama.context_length": 8192,
    "llama.embedding_length": 4096,
    "llama.feed_forward_length": 14336,
    "llama.rope.dimension_count": 128,
    "llama.rope.freq_base": 500000,
    "llama.vocab_size": 128256,
    "tokenizer.ggml.bos_token_id": 128000,
    "tokenizer.ggml.eos_token_id": 128009,
    "tokenizer.ggml.merges": [],
    "tokenizer.ggml.model": "gpt2",
    "tokenizer.ggml.pre": "llama-bpe",
    "tokenizer.ggml.token_type": [],
    "tokenizer.ggml.tokens": []
  }
<!-- gh-comment-id:2187953995 --> @jmorganca commented on GitHub (Jun 25, 2024): Hi all, this should be possible via `/api/show`: ``` "model_info": { "general.architecture": "llama", "general.file_type": 2, "general.parameter_count": 8030261248, "general.quantization_version": 2, "llama.attention.head_count": 32, "llama.attention.head_count_kv": 8, "llama.attention.layer_norm_rms_epsilon": 0.00001, "llama.block_count": 32, "llama.context_length": 8192, "llama.embedding_length": 4096, "llama.feed_forward_length": 14336, "llama.rope.dimension_count": 128, "llama.rope.freq_base": 500000, "llama.vocab_size": 128256, "tokenizer.ggml.bos_token_id": 128000, "tokenizer.ggml.eos_token_id": 128009, "tokenizer.ggml.merges": [], "tokenizer.ggml.model": "gpt2", "tokenizer.ggml.pre": "llama-bpe", "tokenizer.ggml.token_type": [], "tokenizer.ggml.tokens": [] } ```
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 feature request
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#1642
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?