Mirror/ollama4j
1
0
Fork 1
mirror of https://github.com/amithkoujalgi/ollama4j.git synced 2025-05-15 03:47:13 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity

Added new createModel API to make it conform to Ollama's new API - https://github.com/ollama/ollama/blob/main/docs/api.md#create-a-model

Browse Source
This commit is contained in:
Amith Koujalgi 2025-02-17 22:25:25 +05:30
parent e409ff1cf9
commit 23d23c4ad7
5 changed files with 266 additions and 114 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

292
docs/docs/apis-model-management/create-model.md
View File

@ -6,7 +6,7 @@ sidebar_position: 5
This API lets you create a custom model on the Ollama server. This API lets you create a custom model on the Ollama server.
### Create a model from an existing Modelfile in the Ollama server ### Create a custom model from an existing model in the Ollama server
```java title="CreateModel.java" ```java title="CreateModel.java"
import io.github.ollama4j.OllamaAPI; import io.github.ollama4j.OllamaAPI;
@ -19,144 +19,220 @@ public class CreateModel {
OllamaAPI ollamaAPI = new OllamaAPI(host); OllamaAPI ollamaAPI = new OllamaAPI(host);
ollamaAPI.createModelWithFilePath("mario", "/path/to/mario/modelfile/on/ollama-server"); ollamaAPI.createModel(CustomModelRequest.builder().model("mario").from("llama3.2:latest").system("You are Mario from Super Mario Bros.").build());
}
}
```
### Create a model by passing the contents of Modelfile
```java title="CreateModel.java"
public class CreateModel {
public static void main(String[] args) {
String host = "http://localhost:11434/";
OllamaAPI ollamaAPI = new OllamaAPI(host);
ollamaAPI.createModelWithModelFileContents("mario", "FROM llama2\nSYSTEM You are mario from Super Mario Bros.");
} }
} }
``` ```
Once created, you can see it when you use [list models](./list-models) API. Once created, you can see it when you use [list models](./list-models) API.
### Example of a `Modelfile` [Read more](https://github.com/ollama/ollama/blob/main/docs/api.md#create-a-model) about custom model creation and the parameters available for model creation.
``` [//]: # ()
FROM llama2 [//]: # (### Example of a `Modelfile`)
# sets the temperature to 1 [higher is more creative, lower is more coherent]
PARAMETER temperature 1
# sets the context window size to 4096, this controls how many tokens the LLM can use as context to generate the next token
PARAMETER num_ctx 4096
# sets a custom system message to specify the behavior of the chat assistant [//]: # ()
SYSTEM You are Mario from super mario bros, acting as an assistant. [//]: # (```)
```
### Format of the `Modelfile` [//]: # (FROM llama2)
```modelfile [//]: # (# sets the temperature to 1 [higher is more creative, lower is more coherent])
# comment
INSTRUCTION arguments
```
| Instruction | Description | [//]: # (PARAMETER temperature 1)
|-------------------------------------|----------------------------------------------------------------|
| [`FROM`](#from-required) (required) | Defines the base model to use. |
| [`PARAMETER`](#parameter) | Sets the parameters for how Ollama will run the model. |
| [`TEMPLATE`](#template) | The full prompt template to be sent to the model. |
| [`SYSTEM`](#system) | Specifies the system message that will be set in the template. |
| [`ADAPTER`](#adapter) | Defines the (Q)LoRA adapters to apply to the model. |
| [`LICENSE`](#license) | Specifies the legal license. |
#### PARAMETER [//]: # (# sets the context window size to 4096, this controls how many tokens the LLM can use as context to generate the next token)
The `PARAMETER` instruction defines a parameter that can be set when the model is run. [//]: # (PARAMETER num_ctx 4096)
| Parameter | Description | Value Type | Example Usage | [//]: # ()
|----------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------|----------------------| [//]: # (# sets a custom system message to specify the behavior of the chat assistant)
| mirostat | Enable Mirostat sampling for controlling perplexity. (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0) | int | mirostat 0 |
| mirostat_eta | Influences how quickly the algorithm responds to feedback from the generated text. A lower learning rate will result in slower adjustments, while a higher learning rate will make the algorithm more responsive. (Default: 0.1) | float | mirostat_eta 0.1 |
| mirostat_tau | Controls the balance between coherence and diversity of the output. A lower value will result in more focused and coherent text. (Default: 5.0) | float | mirostat_tau 5.0 |
| num_ctx | Sets the size of the context window used to generate the next token. (Default: 2048) | int | num_ctx 4096 |
| num_gqa | The number of GQA groups in the transformer layer. Required for some models, for example it is 8 for llama2:70b | int | num_gqa 1 |
| num_gpu | The number of layers to send to the GPU(s). On macOS it defaults to 1 to enable metal support, 0 to disable. | int | num_gpu 50 |
| num_thread | Sets the number of threads to use during computation. By default, Ollama will detect this for optimal performance. It is recommended to set this value to the number of physical CPU cores your system has (as opposed to the logical number of cores). | int | num_thread 8 |
| repeat_last_n | Sets how far back for the model to look back to prevent repetition. (Default: 64, 0 = disabled, -1 = num_ctx) | int | repeat_last_n 64 |
| repeat_penalty | Sets how strongly to penalize repetitions. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. (Default: 1.1) | float | repeat_penalty 1.1 |
| temperature | The temperature of the model. Increasing the temperature will make the model answer more creatively. (Default: 0.8) | float | temperature 0.7 |
| seed | Sets the random number seed to use for generation. Setting this to a specific number will make the model generate the same text for the same prompt. (Default: 0) | int | seed 42 |
| stop | Sets the stop sequences to use. When this pattern is encountered the LLM will stop generating text and return. Multiple stop patterns may be set by specifying multiple separate `stop` parameters in a modelfile. | string | stop "AI assistant:" |
| tfs_z | Tail free sampling is used to reduce the impact of less probable tokens from the output. A higher value (e.g., 2.0) will reduce the impact more, while a value of 1.0 disables this setting. (default: 1) | float | tfs_z 1 |
| num_predict | Maximum number of tokens to predict when generating text. (Default: 128, -1 = infinite generation, -2 = fill context) | int | num_predict 42 |
| top_k | Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative. (Default: 40) | int | top_k 40 |
| top_p | Works together with top-k. A higher value (e.g., 0.95) will lead to more diverse text, while a lower value (e.g., 0.5) will generate more focused and conservative text. (Default: 0.9) | float | top_p 0.9 |
#### TEMPLATE [//]: # (SYSTEM You are Mario from super mario bros, acting as an assistant.)
`TEMPLATE` of the full prompt template to be passed into the model. It may include (optionally) a system message and a [//]: # (```)
user's prompt. This is used to create a full custom prompt, and syntax may be model specific. You can usually find the
template for a given model in the readme for that model.
#### Template Variables [//]: # ()
[//]: # (### Format of the `Modelfile`)
| Variable | Description | [//]: # ()
|-----------------|---------------------------------------------------------------------------------------------------------------| [//]: # (```modelfile)
| `{{ .System }}` | The system message used to specify custom behavior, this must also be set in the Modelfile as an instruction. |
| `{{ .Prompt }}` | The incoming prompt, this is not specified in the model file and will be set based on input. |
| `{{ .First }}` | A boolean value used to render specific template information for the first generation of a session. |
```modelfile [//]: # (# comment)
TEMPLATE """
{{- if .First }}
### System:
{{ .System }}
{{- end }}
### User: [//]: # (INSTRUCTION arguments)
{{ .Prompt }}
### Response: [//]: # (```)
"""
SYSTEM """<system message>""" [//]: # ()
``` [//]: # (| Instruction | Description |)
### SYSTEM [//]: # (|-------------------------------------|----------------------------------------------------------------|)
The `SYSTEM` instruction specifies the system message to be used in the template, if applicable. [//]: # (| [`FROM`]&#40;#from-required&#41; &#40;required&#41; | Defines the base model to use. |)
```modelfile [//]: # (| [`PARAMETER`]&#40;#parameter&#41; | Sets the parameters for how Ollama will run the model. |)
SYSTEM """<system message>"""
```
### ADAPTER [//]: # (| [`TEMPLATE`]&#40;#template&#41; | The full prompt template to be sent to the model. |)
The `ADAPTER` instruction specifies the LoRA adapter to apply to the base model. The value of this instruction should be [//]: # (| [`SYSTEM`]&#40;#system&#41; | Specifies the system message that will be set in the template. |)
an absolute path or a path relative to the Modelfile and the file must be in a GGML file format. The adapter should be
tuned from the base model otherwise the behaviour is undefined.
```modelfile [//]: # (| [`ADAPTER`]&#40;#adapter&#41; | Defines the &#40;Q&#41;LoRA adapters to apply to the model. |)
ADAPTER ./ollama-lora.bin
```
### LICENSE [//]: # (| [`LICENSE`]&#40;#license&#41; | Specifies the legal license. |)
The `LICENSE` instruction allows you to specify the legal license under which the model used with this Modelfile is [//]: # ()
shared or distributed. [//]: # (#### PARAMETER)
```modelfile [//]: # ()
LICENSE """ [//]: # (The `PARAMETER` instruction defines a parameter that can be set when the model is run.)
<license text>
"""
```
## Notes [//]: # ()
[//]: # (| Parameter | Description | Value Type | Example Usage |)
- the **`Modelfile` is not case sensitive**. In the examples, uppercase instructions are used to make it easier to [//]: # (|----------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------|----------------------|)
distinguish it from arguments.
- Instructions can be in any order. In the examples, the `FROM` instruction is first to keep it easily readable.
Read more about Modelfile: https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md [//]: # (| mirostat | Enable Mirostat sampling for controlling perplexity. &#40;default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0&#41; | int | mirostat 0 |)
[//]: # (| mirostat_eta | Influences how quickly the algorithm responds to feedback from the generated text. A lower learning rate will result in slower adjustments, while a higher learning rate will make the algorithm more responsive. &#40;Default: 0.1&#41; | float | mirostat_eta 0.1 |)
[//]: # (| mirostat_tau | Controls the balance between coherence and diversity of the output. A lower value will result in more focused and coherent text. &#40;Default: 5.0&#41; | float | mirostat_tau 5.0 |)
[//]: # (| num_ctx | Sets the size of the context window used to generate the next token. &#40;Default: 2048&#41; | int | num_ctx 4096 |)
[//]: # (| num_gqa | The number of GQA groups in the transformer layer. Required for some models, for example it is 8 for llama2:70b | int | num_gqa 1 |)
[//]: # (| num_gpu | The number of layers to send to the GPU&#40;s&#41;. On macOS it defaults to 1 to enable metal support, 0 to disable. | int | num_gpu 50 |)
[//]: # (| num_thread | Sets the number of threads to use during computation. By default, Ollama will detect this for optimal performance. It is recommended to set this value to the number of physical CPU cores your system has &#40;as opposed to the logical number of cores&#41;. | int | num_thread 8 |)
[//]: # (| repeat_last_n | Sets how far back for the model to look back to prevent repetition. &#40;Default: 64, 0 = disabled, -1 = num_ctx&#41; | int | repeat_last_n 64 |)
[//]: # (| repeat_penalty | Sets how strongly to penalize repetitions. A higher value &#40;e.g., 1.5&#41; will penalize repetitions more strongly, while a lower value &#40;e.g., 0.9&#41; will be more lenient. &#40;Default: 1.1&#41; | float | repeat_penalty 1.1 |)
[//]: # (| temperature | The temperature of the model. Increasing the temperature will make the model answer more creatively. &#40;Default: 0.8&#41; | float | temperature 0.7 |)
[//]: # (| seed | Sets the random number seed to use for generation. Setting this to a specific number will make the model generate the same text for the same prompt. &#40;Default: 0&#41; | int | seed 42 |)
[//]: # (| stop | Sets the stop sequences to use. When this pattern is encountered the LLM will stop generating text and return. Multiple stop patterns may be set by specifying multiple separate `stop` parameters in a modelfile. | string | stop "AI assistant:" |)
[//]: # (| tfs_z | Tail free sampling is used to reduce the impact of less probable tokens from the output. A higher value &#40;e.g., 2.0&#41; will reduce the impact more, while a value of 1.0 disables this setting. &#40;default: 1&#41; | float | tfs_z 1 |)
[//]: # (| num_predict | Maximum number of tokens to predict when generating text. &#40;Default: 128, -1 = infinite generation, -2 = fill context&#41; | int | num_predict 42 |)
[//]: # (| top_k | Reduces the probability of generating nonsense. A higher value &#40;e.g. 100&#41; will give more diverse answers, while a lower value &#40;e.g. 10&#41; will be more conservative. &#40;Default: 40&#41; | int | top_k 40 |)
[//]: # (| top_p | Works together with top-k. A higher value &#40;e.g., 0.95&#41; will lead to more diverse text, while a lower value &#40;e.g., 0.5&#41; will generate more focused and conservative text. &#40;Default: 0.9&#41; | float | top_p 0.9 |)
[//]: # ()
[//]: # (#### TEMPLATE)
[//]: # ()
[//]: # (`TEMPLATE` of the full prompt template to be passed into the model. It may include &#40;optionally&#41; a system message and a)
[//]: # (user's prompt. This is used to create a full custom prompt, and syntax may be model specific. You can usually find the)
[//]: # (template for a given model in the readme for that model.)
[//]: # ()
[//]: # (#### Template Variables)
[//]: # ()
[//]: # (| Variable | Description |)
[//]: # (|-----------------|---------------------------------------------------------------------------------------------------------------|)
[//]: # (| `{{ .System }}` | The system message used to specify custom behavior, this must also be set in the Modelfile as an instruction. |)
[//]: # (| `{{ .Prompt }}` | The incoming prompt, this is not specified in the model file and will be set based on input. |)
[//]: # (| `{{ .First }}` | A boolean value used to render specific template information for the first generation of a session. |)
[//]: # ()
[//]: # (```modelfile)
[//]: # (TEMPLATE """)
[//]: # ({{- if .First }})
[//]: # (### System:)
[//]: # ({{ .System }})
[//]: # ({{- end }})
[//]: # ()
[//]: # (### User:)
[//]: # ({{ .Prompt }})
[//]: # ()
[//]: # (### Response:)
[//]: # (""")
[//]: # ()
[//]: # (SYSTEM """<system message>""")
[//]: # (```)
[//]: # ()
[//]: # (### SYSTEM)
[//]: # ()
[//]: # (The `SYSTEM` instruction specifies the system message to be used in the template, if applicable.)
[//]: # ()
[//]: # (```modelfile)
[//]: # (SYSTEM """<system message>""")
[//]: # (```)
[//]: # ()
[//]: # (### ADAPTER)
[//]: # ()
[//]: # (The `ADAPTER` instruction specifies the LoRA adapter to apply to the base model. The value of this instruction should be)
[//]: # (an absolute path or a path relative to the Modelfile and the file must be in a GGML file format. The adapter should be)
[//]: # (tuned from the base model otherwise the behaviour is undefined.)
[//]: # ()
[//]: # (```modelfile)
[//]: # (ADAPTER ./ollama-lora.bin)
[//]: # (```)
[//]: # ()
[//]: # (### LICENSE)
[//]: # ()
[//]: # (The `LICENSE` instruction allows you to specify the legal license under which the model used with this Modelfile is)
[//]: # (shared or distributed.)
[//]: # ()
[//]: # (```modelfile)
[//]: # (LICENSE """)
[//]: # (<license text>)
[//]: # (""")
[//]: # (```)
[//]: # ()
[//]: # (## Notes)
[//]: # ()
[//]: # (- the **`Modelfile` is not case sensitive**. In the examples, uppercase instructions are used to make it easier to)
[//]: # ( distinguish it from arguments.)
[//]: # (- Instructions can be in any order. In the examples, the `FROM` instruction is first to keep it easily readable.)
[//]: # ()
[//]: # (Read more about Modelfile: https://github.com/jmorganca/ollama/blob/main/docs/modelfile.md)

31
src/main/java/io/github/ollama4j/OllamaAPI.java
View File

@ -391,6 +391,7 @@ public class OllamaAPI {
* @throws InterruptedException if the operation is interrupted * @throws InterruptedException if the operation is interrupted
* @throws URISyntaxException if the URI for the request is malformed * @throws URISyntaxException if the URI for the request is malformed
*/ */
@Deprecated
public void createModelWithFilePath(String modelName, String modelFilePath) throws IOException, InterruptedException, OllamaBaseException, URISyntaxException { public void createModelWithFilePath(String modelName, String modelFilePath) throws IOException, InterruptedException, OllamaBaseException, URISyntaxException {
String url = this.host + "/api/create"; String url = this.host + "/api/create";
String jsonData = new CustomModelFilePathRequest(modelName, modelFilePath).toString(); String jsonData = new CustomModelFilePathRequest(modelName, modelFilePath).toString();
@ -423,6 +424,7 @@ public class OllamaAPI {
* @throws InterruptedException if the operation is interrupted * @throws InterruptedException if the operation is interrupted
* @throws URISyntaxException if the URI for the request is malformed * @throws URISyntaxException if the URI for the request is malformed
*/ */
@Deprecated
public void createModelWithModelFileContents(String modelName, String modelFileContents) throws IOException, InterruptedException, OllamaBaseException, URISyntaxException { public void createModelWithModelFileContents(String modelName, String modelFileContents) throws IOException, InterruptedException, OllamaBaseException, URISyntaxException {
String url = this.host + "/api/create"; String url = this.host + "/api/create";
String jsonData = new CustomModelFileContentsRequest(modelName, modelFileContents).toString(); String jsonData = new CustomModelFileContentsRequest(modelName, modelFileContents).toString();
@ -442,6 +444,35 @@ public class OllamaAPI {
} }
} }
/**
* Create a custom model. Read more about custom model creation <a
* href="https://github.com/ollama/ollama/blob/main/docs/api.md#create-a-model">here</a>.
*
* @param customModelRequest custom model spec
* @throws OllamaBaseException if the response indicates an error status
* @throws IOException if an I/O error occurs during the HTTP request
* @throws InterruptedException if the operation is interrupted
* @throws URISyntaxException if the URI for the request is malformed
*/
public void createModel(CustomModelRequest customModelRequest) throws IOException, InterruptedException, OllamaBaseException, URISyntaxException {
String url = this.host + "/api/create";
String jsonData = customModelRequest.toString();
HttpRequest request = getRequestBuilderDefault(new URI(url)).header("Accept", "application/json").header("Content-Type", "application/json").POST(HttpRequest.BodyPublishers.ofString(jsonData, StandardCharsets.UTF_8)).build();
HttpClient client = HttpClient.newHttpClient();
HttpResponse<String> response = client.send(request, HttpResponse.BodyHandlers.ofString());
int statusCode = response.statusCode();
String responseString = response.body();
if (statusCode != 200) {
throw new OllamaBaseException(statusCode + " - " + responseString);
}
if (responseString.contains("error")) {
throw new OllamaBaseException(responseString);
}
if (verbose) {
logger.info(responseString);
}
}
/** /**
* Delete a model from Ollama server. * Delete a model from Ollama server.
* *

2
src/main/java/io/github/ollama4j/models/embeddings/OllamaEmbedRequestBuilder.java
View File

@ -5,7 +5,7 @@ import io.github.ollama4j.utils.Options;
import java.util.List; import java.util.List;
/** /**
* Builderclass to easily create Requests for Embedding models using ollama. * Builder class to easily create Requests for Embedding models using ollama.
*/ */
public class OllamaEmbedRequestBuilder { public class OllamaEmbedRequestBuilder {

45
src/main/java/io/github/ollama4j/models/request/CustomModelRequest.java Normal file
View File

@ -0,0 +1,45 @@
package io.github.ollama4j.models.request;
import static io.github.ollama4j.utils.Utils.getObjectMapper;
import com.fasterxml.jackson.core.JsonProcessingException;
import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.Data;
import lombok.AllArgsConstructor;
import lombok.Builder;
import java.util.List;
import java.util.Map;
@Data
@AllArgsConstructor
@Builder
public class CustomModelRequest {
private String model;
private String from;
private Map<String, String> files;
private Map<String, String> adapters;
private String template;
private Object license; // Using Object to handle both String and List<String>
private String system;
private Map<String, Object> parameters;
private List<Object> messages;
private Boolean stream;
private Boolean quantize;
public CustomModelRequest() {
this.stream = true;
this.quantize = false;
}
@Override
public String toString() {
try {
return getObjectMapper().writerWithDefaultPrettyPrinter().writeValueAsString(this);
} catch (JsonProcessingException e) {
throw new RuntimeException(e);
}
}
}

10
src/test/java/io/github/ollama4j/unittests/TestMockedAPIs.java
View File

@ -6,6 +6,7 @@ import io.github.ollama4j.exceptions.RoleNotFoundException;
import io.github.ollama4j.models.chat.OllamaChatMessageRole; import io.github.ollama4j.models.chat.OllamaChatMessageRole;
import io.github.ollama4j.models.embeddings.OllamaEmbedRequestModel; import io.github.ollama4j.models.embeddings.OllamaEmbedRequestModel;
import io.github.ollama4j.models.embeddings.OllamaEmbedResponseModel; import io.github.ollama4j.models.embeddings.OllamaEmbedResponseModel;
import io.github.ollama4j.models.request.CustomModelRequest;
import io.github.ollama4j.models.response.ModelDetail; import io.github.ollama4j.models.response.ModelDetail;
import io.github.ollama4j.models.response.OllamaAsyncResultStreamer; import io.github.ollama4j.models.response.OllamaAsyncResultStreamer;
import io.github.ollama4j.models.response.OllamaResult; import io.github.ollama4j.models.response.OllamaResult;
@ -52,12 +53,11 @@ class TestMockedAPIs {
@Test @Test
void testCreateModel() { void testCreateModel() {
OllamaAPI ollamaAPI = Mockito.mock(OllamaAPI.class); OllamaAPI ollamaAPI = Mockito.mock(OllamaAPI.class);
String model = OllamaModelType.LLAMA2; CustomModelRequest customModelRequest = CustomModelRequest.builder().model("mario").from("llama3.2:latest").system("You are Mario from Super Mario Bros.").build();
String modelFilePath = "FROM llama2\nSYSTEM You are mario from Super Mario Bros.";
try { try {
doNothing().when(ollamaAPI).createModelWithModelFileContents(model, modelFilePath); doNothing().when(ollamaAPI).createModel(customModelRequest);
ollamaAPI.createModelWithModelFileContents(model, modelFilePath); ollamaAPI.createModel(customModelRequest);
verify(ollamaAPI, times(1)).createModelWithModelFileContents(model, modelFilePath); verify(ollamaAPI, times(1)).createModel(customModelRequest);
} catch (IOException | OllamaBaseException | InterruptedException | URISyntaxException e) { } catch (IOException | OllamaBaseException | InterruptedException | URISyntaxException e) {
throw new RuntimeException(e); throw new RuntimeException(e);
} }