Skip to content

Model repetitions on Intel CPU with qwen3-1.7b-generic-cpu:2 #896

Description

@natke

Model repeats after a few round of inputs. Here is the chat session

oundry run qwen3-1.7b-generic-cpu:2
■ Chatting model (qwen3-1.7b)
■ note: Type /help to see chat commands, /exit to quit.
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /help
╭────────────────────────────┬────────────────────────────────────────────────────────────────────────────╮
│ Command                    │ Description                                                                │
├────────────────────────────┼────────────────────────────────────────────────────────────────────────────┤
│ /help (alias /?)           │ Show available slash commands.                                             │
│ /exit (alias /quit)        │ Leave the chat session.                                                    │
│ /clear                     │ Clear the conversation history (keeps the system prompt).                  │
│ /history                   │ Show the conversation history.                                             │
│ /copy                      │ Copy the previous assistant response to the clipboard.                     │
│ /add                       │ Append the contents of <file> as a user message.                           │
│ /context                   │ Show context-window usage and per-message token estimates.                 │
│ /system                    │ Set or replace the system prompt.                                          │
│ /temperature (alias /temp) │ Show or set the sampling temperature ([0.0, 1.0]; 'default' to clear).     │
│ /max-tokens                │ Show or set the per-turn token cap (positive integer; 'default' to clear). │
│ /voice                     │ Speak a prompt with press-and-hold Space.                                  │
╰────────────────────────────┴────────────────────────────────────────────────────────────────────────────╯
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /context
■ note: Model: qwen3-1.7b  Context: 0 / 40960 tokens (0%)
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /max-tokens
■ note: Max tokens: default
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /max-tokens:1000
● error: Unknown slash command '/max-tokens:1000'.
Hint: Type /help to list available commands.
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /max-tokens=1000
● error: Unknown slash command '/max-tokens=1000'.
Hint: Type /help to list available commands.
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /max-tokens 1000
● success: Max tokens set to 1000.
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ data data data data data data data data data
<think>
Okay, the user has sent a bunch of "data data data..." messages. They probably want me to respond to this repetitive input.

First, I need to understand the context. The user might be testing if I can handle repetitive messages. Alternatively, they might be trying to get a response from the system.

Since the user has sent a lot of "data data data..." messages, it's possible that they are trying to get a response from the system. However, the system might not be able to process the repetitive messages, and thus the system might not be able to respond to the user's input.

In summary, the user has sent a lot of "data data data..." messages, and the system might not be able to process the repetitive messages, and thus the system might not be able to respond to the user's input.
</think>

The repeated "data data data..." messages suggest a pattern of repetition. If this is a test or a mistake, I’d be happy to help with any specific questions you might have. 😊
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /context
■ note: Model: qwen3-1.7b  Context: 256 / 40960 tokens (1%)
╭───┬───────────┬────────╮
│ # │ Role      │ Tokens │
├───┼───────────┼────────┤
│ 1 │ user      │ 11     │
│ 2 │ assistant │ 245    │
╰───┴───────────┴────────╯
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ Hello
<think>
Okay, the user just said "Hello". I need to respond appropriately. Since the user is using a repeated message, maybe they're testing the system or just trying to get a response.

In this case, the user is likely a beginner trying to get a response. The user might be testing if the system can handle repeated messages.

So, the response should be friendly and acknowledge the greeting. Maybe add a brief message to keep the conversation going.

In summary, the user's message is "Hello". The response should be friendly and acknowledge the greeting, possibly adding a brief message to keep the conversation going.
</think>

Hello! 😊

I hope you're doing well. If you have any questions or need assistance, feel free to ask! 😊
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /context
■ note: Model: qwen3-1.7b  Context: 443 / 40960 tokens (1%)
╭───┬───────────┬────────╮
│ # │ Role      │ Tokens │
├───┼───────────┼────────┤
│ 1 │ user      │ 11     │
│ 2 │ assistant │ 245    │
│ 3 │ user      │ 2      │
│ 4 │ assistant │ 185    │
╰───┴───────────┴────────╯
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ Thank you, what is your name?
<think>
Okay, the user just said "Thank you, what is your name?" and I need to respond.

First, I should acknowledge the thanks and the question about my name. The user is probably testing if I can handle such a question, or maybe they're just curious.

Since the user is asking about my name, I should respond in a friendly and helpful manner. Maybe add a brief message to keep the conversation going.

Also, considering the user's repeated messages, it's possible that they're testing the system or just curious. So, the response should be friendly and helpful, regardless of the user's intent.

In summary, the user is asking about my name, and I need to respond in a friendly and helpful manner.
</think>

Hello! 😊

I hope you're doing well. If you have any questions or need assistance, feel free to ask! 😊
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /context
■ note: Model: qwen3-1.7b  Context: 655 / 40960 tokens (2%)
╭───┬───────────┬────────╮
│ # │ Role      │ Tokens │
├───┼───────────┼────────┤
│ 1 │ user      │ 11     │
│ 2 │ assistant │ 245    │
│ 3 │ user      │ 2      │
│ 4 │ assistant │ 185    │
│ 5 │ user      │ 8      │
│ 6 │ assistant │ 204    │
╰───┴───────────┴────────╯
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ Write a python function to calculate the volume of a house, given height, length, and width of all rooms in the house
<think>
Okay, the user wants a Python function to calculate the volume of a house. The parameters given are height, length, and width of all rooms in the house.

First, I need to understand what exactly the user is asking for. The volume of a house, given the height, length, and width of all rooms. So, the volume would be the sum of the volumes of all the rooms in the house.

So, the function needs to take the height, length, and width of all rooms in the house. Wait, the parameters given are height, length, and width of all rooms in the house. So, the user is probably referring to the height of the house, and the length and width of all the rooms in the house.

So, the volume of the house would be the sum of the volumes of all the rooms in the house.

Therefore, the function would need to take the height, length, and width of all rooms in the house, and then compute the volume as the sum of the volumes of all the rooms.

So, the Python function would look something like this:

def calculate_house_volume(height, length, width):
    # Calculate the volume of each room
    # Assuming all rooms are rectangular
    # For simplicity, we assume that the length and width are the same for all rooms
    # So, the volume of each room is length * width
    # Since all rooms are rectangular and have the same length and width, the volume of the house is (height + length + width) * (length * width)
    # Wait, this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume of the house would be the sum of the volumes of all the rooms in the house. If all rooms are rectangular and have the same length and width, then the volume of the house is (height + length + width) * (length * width)
    # But this is not correct. The volume
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯ /context
■ note: Model: qwen3-1.7b  Context: 1763 / 40960 tokens (4%)
╭───┬───────────┬────────╮
│ # │ Role      │ Tokens │
├───┼───────────┼────────┤
│ 1 │ user      │ 11     │
│ 2 │ assistant │ 245    │
│ 3 │ user      │ 2      │
│ 4 │ assistant │ 185    │
│ 5 │ user      │ 8      │
│ 6 │ assistant │ 204    │
│ 7 │ user      │ 30     │
│ 8 │ assistant │ 1078   │
╰───┴───────────┴────────╯
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
❯

Machine details

foundry status
╭──────────────┬────────────────────┬─────────────────────────────────────────────────────────╮
│ Section      │ Metric             │ Value                                                   │
├──────────────┼────────────────────┼─────────────────────────────────────────────────────────┤
│ System       │ OS                 │ Microsoft Windows 10.0.26655                            │
│ System       │ Architecture       │ X64                                                     │
│ System       │ CPU                │ 12th Gen Intel(R) Core(TM) i7-1265U (12 logical cores)  │
│ System       │ GPU                │ Intel Corporation Intel(R) Iris(R) Xe Graphics (2.0 GB) │
│ System       │ RAM                │ 11.6 GB available / 31.8 GB total                       │
│ System       │ Disk               │ 2.1 GB free / 952.5 GB total (C:\)                      │
│ System       │ .NET               │ .NET 9.0.18                                             │
│ Service      │ State              │ Ready                                                   │
│ Service      │ Web URLs           │ http://127.0.0.1:53592                                  │
│ Service      │ CLI version        │ 0.10.2                                                  │
│ Models       │ Available variants │ 100                                                     │
│ Models       │ Loaded             │ 4                                                       │
│ Models       │ Cached             │ 8                                                       │
│ Connectivity │ Local service      │ Reachable                                               │
│ Connectivity │ Model registry     │ Reachable                                               │
│ Warnings     │ Status             │ No warnings.                                            │
╰──────────────┴────────────────────┴─────────────────────────────────────────────────────────╯

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions