ollama/llama3.2 maximum sequence length

[AceCodePt]

graph_config = {
    "llm": {
        "model": "ollama/llama3.2",
        "temperature": 0,
        "model_tokens":10000,
        "format": "json",  # Ollama needs the format to be specified explicitly
        # "base_url": "http://localhost:11434", # set ollama URL arbitrarily
    },
    "verbose": True,
    "headless": False
}

Result with:
Token indices sequence length is longer than the specified maximum sequence length for this model (1385 > 1024). Running this sequence through the model will result in indexing errors

smart_scraper_graph = SmartScraperGraph(
    prompt="List all links",
    source="https://www.google.com/search?q=why+it+doesn%27t+work",
    schema=Properties,
    config=graph_config
)

Originally posted by @AceCodePt in #875 (comment)

[dosubot]

Hey there, @AceCodePt! 👋 I'm here to help with any bugs, questions, or contributions you may have while waiting for a human maintainer. Let's tackle this together!

There is a similar closed issue regarding the sequence length problem with the ollama/llama3.2 model. The issue was resolved in a pull request that made the model_tokens dictionary key accessible for all models. This fix was included in version 1.27.0-beta.13 and later confirmed to be resolved in version 1.28.0. Updating to version 1.28.0 or later should resolve the issue ^[1][2].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[PeriniM]

Hey @AceCodePt this issue has just been fixed in v1.36.0, try it out! #856