From cc0cca539f118cf232337a47cb85f6eb19952ce3 Mon Sep 17 00:00:00 2001 From: jjjuk Date: Thu, 24 Jul 2025 03:53:23 +0300 Subject: [PATCH 1/5] docs: Tool calling support and Context size warning Signed-off-by: jjjuk --- documentation/docs/getting-started/providers.md | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/documentation/docs/getting-started/providers.md b/documentation/docs/getting-started/providers.md index fac23ce56322..532d2e8a9ae8 100644 --- a/documentation/docs/getting-started/providers.md +++ b/documentation/docs/getting-started/providers.md @@ -305,8 +305,10 @@ Ollama and Ramalama are both options to provide local LLMs, each which requires 1. [Download Ollama](https://ollama.com/download). 2. Run any [model supporting tool-calling](https://ollama.com/search?c=tools): -:::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. +:::warning Tool calling support and Context size +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. + +Also, ensure that the `OLLAMA_CONTEXT_LENGTH` environment variable is set high enough to accommodate the instructions required for Goose and extensions. If Goose uses extensions incorrectly or ignores `.goosehits` file, the context length is likely too low. Refer to the [Ollama documentation](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size) for more details. ::: Example: @@ -402,15 +404,17 @@ If you're running Ollama on a different server, you'll have to set `OLLAMA_HOST= 1. [Download Ramalama](https://github.com/containers/ramalama?tab=readme-ov-file#install). 2. Run any Ollama [model supporting tool-calling](https://ollama.com/search?c=tools) or [GGUF format HuggingFace Model](https://huggingface.co/search/full-text?q=%22tools+support%22+%2B+%22gguf%22&type=model) : -:::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. +:::warning Tool calling support and Context size +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. + +Using `ramalama serve`, ensure that the `--ctx-size, -c` option is set high enough to accommodate the instructions required for Goose and extensions. If Goose uses extensions incorrectly or ignores `.goosehits` file, the context length is likely too low. Refer to the [Ramalama documentation](https://github.com/containers/ramalama/blob/main/docs/ramalama-serve.1.md#--ctx-size--c) for more details. ::: Example: ```sh # NOTE: the --runtime-args="--jinja" flag is required for Ramalama to work with the Goose Ollama provider. -ramalama serve --runtime-args="--jinja" ollama://qwen2.5 +ramalama serve --runtime-args="--jinja" --ctx-size=8192 ollama://qwen2.5 ``` 3. In a separate terminal window, configure with Goose: From 7440e1ea1f14b9002b30256f4125a96d08902b13 Mon Sep 17 00:00:00 2001 From: angiejones Date: Wed, 23 Jul 2025 18:56:58 -0500 Subject: [PATCH 2/5] cleaning up --- .../docs/getting-started/providers.md | 21 ++++++++++++------- 1 file changed, 13 insertions(+), 8 deletions(-) diff --git a/documentation/docs/getting-started/providers.md b/documentation/docs/getting-started/providers.md index 532d2e8a9ae8..742253df4429 100644 --- a/documentation/docs/getting-started/providers.md +++ b/documentation/docs/getting-started/providers.md @@ -305,10 +305,8 @@ Ollama and Ramalama are both options to provide local LLMs, each which requires 1. [Download Ollama](https://ollama.com/download). 2. Run any [model supporting tool-calling](https://ollama.com/search?c=tools): -:::warning Tool calling support and Context size -Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. - -Also, ensure that the `OLLAMA_CONTEXT_LENGTH` environment variable is set high enough to accommodate the instructions required for Goose and extensions. If Goose uses extensions incorrectly or ignores `.goosehits` file, the context length is likely too low. Refer to the [Ollama documentation](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size) for more details. +:::warning Tool calling support +Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. ::: Example: @@ -399,15 +397,17 @@ If you're running Ollama on a different server, you'll have to set `OLLAMA_HOST= └ Configuration saved successfully ``` +:::tip Context Length +If you notice that Goose is having trouble using extensions or is ignoring [.goosehints](/docs/guides/using-goosehints), it is likely that the model's default context length of 4096 tokens is too low. Set the `OLLAMA_CONTEXT_LENGTH` environment variable to a [higher value](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size). +::: + #### Ramalama 1. [Download Ramalama](https://github.com/containers/ramalama?tab=readme-ov-file#install). 2. Run any Ollama [model supporting tool-calling](https://ollama.com/search?c=tools) or [GGUF format HuggingFace Model](https://huggingface.co/search/full-text?q=%22tools+support%22+%2B+%22gguf%22&type=model) : -:::warning Tool calling support and Context size -Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. - -Using `ramalama serve`, ensure that the `--ctx-size, -c` option is set high enough to accommodate the instructions required for Goose and extensions. If Goose uses extensions incorrectly or ignores `.goosehits` file, the context length is likely too low. Refer to the [Ramalama documentation](https://github.com/containers/ramalama/blob/main/docs/ramalama-serve.1.md#--ctx-size--c) for more details. +:::warning Tool calling support +Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. ::: Example: @@ -497,6 +497,11 @@ For the Ollama provider, if you don't provide a host, we set it to `localhost:11 └ Configuration saved successfully ``` +:::tip Context Length +If you notice that Goose is having trouble using extensions or is ignoring [.goosehints](/docs/guides/using-goosehints), it is likely that the model's default context length of 2048 tokens is too low. Use `ramalama serve` to set the `--ctx-size, -c` option to a [higher value](https://github.com/containers/ramalama/blob/main/docs/ramalama-serve.1.md#--ctx-size--c). +::: + + ### DeepSeek-R1 Ollama provides open source LLMs, such as `DeepSeek-r1`, that you can install and run locally. From 083f4aa34e5949ea54bb6277d81ccc160c6df215 Mon Sep 17 00:00:00 2001 From: angiejones Date: Wed, 23 Jul 2025 19:00:49 -0500 Subject: [PATCH 3/5] restoring warning labels --- documentation/docs/getting-started/providers.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/documentation/docs/getting-started/providers.md b/documentation/docs/getting-started/providers.md index 742253df4429..a3c757a8b212 100644 --- a/documentation/docs/getting-started/providers.md +++ b/documentation/docs/getting-started/providers.md @@ -305,7 +305,7 @@ Ollama and Ramalama are both options to provide local LLMs, each which requires 1. [Download Ollama](https://ollama.com/download). 2. Run any [model supporting tool-calling](https://ollama.com/search?c=tools): -:::warning Tool calling support +:::warning Limited Support for models without tool calling Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. ::: @@ -406,7 +406,7 @@ If you notice that Goose is having trouble using extensions or is ignoring [.goo 1. [Download Ramalama](https://github.com/containers/ramalama?tab=readme-ov-file#install). 2. Run any Ollama [model supporting tool-calling](https://ollama.com/search?c=tools) or [GGUF format HuggingFace Model](https://huggingface.co/search/full-text?q=%22tools+support%22+%2B+%22gguf%22&type=model) : -:::warning Tool calling support +:::warning Limited Support for models without tool calling Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. ::: From b8deabfedbcc87398d977d38499e1836c9e1b9ae Mon Sep 17 00:00:00 2001 From: angiejones Date: Wed, 23 Jul 2025 20:06:53 -0500 Subject: [PATCH 4/5] additional cleanup --- documentation/docs/getting-started/providers.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/documentation/docs/getting-started/providers.md b/documentation/docs/getting-started/providers.md index a3c757a8b212..0afe9322355f 100644 --- a/documentation/docs/getting-started/providers.md +++ b/documentation/docs/getting-started/providers.md @@ -306,7 +306,7 @@ Ollama and Ramalama are both options to provide local LLMs, each which requires 2. Run any [model supporting tool-calling](https://ollama.com/search?c=tools): :::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. ::: Example: @@ -407,7 +407,7 @@ If you notice that Goose is having trouble using extensions or is ignoring [.goo 2. Run any Ollama [model supporting tool-calling](https://ollama.com/search?c=tools) or [GGUF format HuggingFace Model](https://huggingface.co/search/full-text?q=%22tools+support%22+%2B+%22gguf%22&type=model) : :::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it (e.g. `DeepSeek-r1`) can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). As an alternative, you can use a [custom DeepSeek-r1 model](/docs/getting-started/providers#deepseek-r1) we've made specifically for Goose. +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. ::: Example: From bbcf69d1808928f49045f78e585c53cab6c9bcff Mon Sep 17 00:00:00 2001 From: angiejones Date: Wed, 23 Jul 2025 20:10:48 -0500 Subject: [PATCH 5/5] add link back --- documentation/docs/getting-started/providers.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/documentation/docs/getting-started/providers.md b/documentation/docs/getting-started/providers.md index 0afe9322355f..e787b9cb68ec 100644 --- a/documentation/docs/getting-started/providers.md +++ b/documentation/docs/getting-started/providers.md @@ -306,7 +306,7 @@ Ollama and Ramalama are both options to provide local LLMs, each which requires 2. Run any [model supporting tool-calling](https://ollama.com/search?c=tools): :::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). ::: Example: @@ -407,7 +407,7 @@ If you notice that Goose is having trouble using extensions or is ignoring [.goo 2. Run any Ollama [model supporting tool-calling](https://ollama.com/search?c=tools) or [GGUF format HuggingFace Model](https://huggingface.co/search/full-text?q=%22tools+support%22+%2B+%22gguf%22&type=model) : :::warning Limited Support for models without tool calling -Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose extensions must be disabled. +Goose extensively uses tool calling, so models without it can only do chat completion. If using models without tool calling, all Goose [extensions must be disabled](/docs/getting-started/using-extensions#enablingdisabling-extensions). ::: Example: