dereklck commited on
Commit
585744d
·
verified ·
1 Parent(s): 1c98246

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +27 -8
README.md CHANGED
@@ -22,6 +22,8 @@ language:
22
 
23
  This Llama-based model was fine-tuned to generate `kubectl` commands based on user descriptions. It was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Hugging Face's TRL library.
24
 
 
 
25
  ## Model Details
26
 
27
  ### Purpose
@@ -51,6 +53,8 @@ While the model aims to generate accurate commands, there might be occasional ha
51
  - Validate the generated commands before execution.
52
  - Use the model as a guide rather than an absolute source.
53
 
 
 
54
  ## Usage
55
 
56
  ### Prompt Template
@@ -73,6 +77,8 @@ You are an AI assistant that generates kubectl commands based on the user's desc
73
  - **Temperature:** `0.3` (controls the randomness; lower values make outputs more deterministic)
74
  - **Stop Token:** `</s>` (indicates the end of the assistant's response)
75
 
 
 
76
  ### Examples
77
 
78
  #### Example 1: Generating a Command
@@ -90,6 +96,8 @@ kubectl get pods -n kube-system
90
  </s>
91
  ```
92
 
 
 
93
  #### Example 2: Handling Ambiguity
94
 
95
  **Instruction:**
@@ -105,6 +113,8 @@ Could you please specify the name of the deployment you want to restart?
105
  </s>
106
  ```
107
 
 
 
108
  #### Example 3: Providing Explanations
109
 
110
  **Instruction:**
@@ -120,6 +130,8 @@ A Kubernetes Service is an abstraction that defines a logical set of Pods and a
120
  </s>
121
  ```
122
 
 
 
123
  ## Deployment with Ollama
124
 
125
  ### Prerequisites
@@ -129,9 +141,9 @@ A Kubernetes Service is an abstraction that defines a logical set of Pods and a
129
 
130
  ### Steps
131
 
132
- 1. **Create the Model**
133
 
134
- Save the following content as `modelfile_kubectl_operator`:
135
 
136
  ```plaintext
137
  FROM kubectl_operator.Q8_0.gguf
@@ -153,13 +165,17 @@ A Kubernetes Service is an abstraction that defines a logical set of Pods and a
153
  """
154
  ```
155
 
156
- Then, create the model using Ollama:
 
 
157
 
158
  ```bash
159
- ollama create kubectl_operator -f modelfile_kubectl_operator
160
  ```
161
 
162
- 2. **Run the Model**
 
 
163
 
164
  Start interacting with your model:
165
 
@@ -167,10 +183,9 @@ A Kubernetes Service is an abstraction that defines a logical set of Pods and a
167
  ollama run kubectl_operator
168
  ```
169
 
170
- ### Notes
171
 
172
- - Replace `kubectl_operator` with your desired model name if different.
173
- - The `ollama run` command initializes the model and waits for user input based on the prompt template.
174
 
175
  ## Limitations and Considerations
176
 
@@ -178,9 +193,13 @@ A Kubernetes Service is an abstraction that defines a logical set of Pods and a
178
  - **Hallucinations:** In rare cases, the model might generate irrelevant information. If the response seems off-topic, consider rephrasing your instruction.
179
  - **Security:** Be cautious when executing generated commands, especially in production environments.
180
 
 
 
181
  ## Feedback and Contributions
182
 
183
  We welcome any comments or participation to improve the model and dataset. If you encounter issues or have suggestions for improvement:
184
 
185
  - **GitHub:** [Unsloth Repository](https://github.com/unslothai/unsloth)
186
  - **Contact:** Reach out to the developer, **dereklck**, for further assistance.
 
 
 
22
 
23
  This Llama-based model was fine-tuned to generate `kubectl` commands based on user descriptions. It was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Hugging Face's TRL library.
24
 
25
+ ---
26
+
27
  ## Model Details
28
 
29
  ### Purpose
 
53
  - Validate the generated commands before execution.
54
  - Use the model as a guide rather than an absolute source.
55
 
56
+ ---
57
+
58
  ## Usage
59
 
60
  ### Prompt Template
 
77
  - **Temperature:** `0.3` (controls the randomness; lower values make outputs more deterministic)
78
  - **Stop Token:** `</s>` (indicates the end of the assistant's response)
79
 
80
+ ---
81
+
82
  ### Examples
83
 
84
  #### Example 1: Generating a Command
 
96
  </s>
97
  ```
98
 
99
+ ---
100
+
101
  #### Example 2: Handling Ambiguity
102
 
103
  **Instruction:**
 
113
  </s>
114
  ```
115
 
116
+ ---
117
+
118
  #### Example 3: Providing Explanations
119
 
120
  **Instruction:**
 
130
  </s>
131
  ```
132
 
133
+ ---
134
+
135
  ## Deployment with Ollama
136
 
137
  ### Prerequisites
 
141
 
142
  ### Steps
143
 
144
+ 1. **Create the Modelfile**
145
 
146
+ Save the following content as a file named `modelfile`:
147
 
148
  ```plaintext
149
  FROM kubectl_operator.Q8_0.gguf
 
165
  """
166
  ```
167
 
168
+ 2. **Create the Model with Ollama**
169
+
170
+ Open your terminal and run the following command to create the model:
171
 
172
  ```bash
173
+ ollama create kubectl_operator -f modelfile
174
  ```
175
 
176
+ This command tells Ollama to create a new model named `kubectl_operator` using the configuration specified in `modelfile`.
177
+
178
+ 3. **Run the Model**
179
 
180
  Start interacting with your model:
181
 
 
183
  ollama run kubectl_operator
184
  ```
185
 
186
+ This will initiate the model and prompt you for input based on the template provided.
187
 
188
+ ---
 
189
 
190
  ## Limitations and Considerations
191
 
 
193
  - **Hallucinations:** In rare cases, the model might generate irrelevant information. If the response seems off-topic, consider rephrasing your instruction.
194
  - **Security:** Be cautious when executing generated commands, especially in production environments.
195
 
196
+ ---
197
+
198
  ## Feedback and Contributions
199
 
200
  We welcome any comments or participation to improve the model and dataset. If you encounter issues or have suggestions for improvement:
201
 
202
  - **GitHub:** [Unsloth Repository](https://github.com/unslothai/unsloth)
203
  - **Contact:** Reach out to the developer, **dereklck**, for further assistance.
204
+
205
+ ---