[Bug]: when i use qwen embedding model to index, i get a problem

[zzzengzhe]

Do you need to file an issue?

• I have searched the existing issues and this bug is not already filed.
• My model is hosted on OpenAI or Azure. If not, please look at the "model providers" issue and don't file a new one here.
• I believe this is a legitimate bug, not just a question. If this is a question, please use the Discussions area.

Describe the bug

when i use the command "graphrag index --root ./graphrag",i got the wrong message as flow
Failed to validate embedding model (default_embedding_model) params litellm.BadRequestError: OpenAIException - Error code: 400 - {'error': {'message': "'encoding_format' only support with [float, base64]", 'type': 'invalid_request_error', 'param': None, 'code': None}, 'request_id': '42b44c10-cbe7-9a54-9f07-f831683a27fc'}

settings.yaml

This config file contains required core defaults that must be set, along with a handful of common optional settings.

For a full list of available settings, see https://microsoft.github.io/graphrag/config/yaml/

LLM settings

There are a number of settings to tune the threading and token limits for LLM calls - check the docs.

models:
default_chat_model:
type: chat
model_provider: openai
auth_type: api_key # or azure_managed_identity
api_key: ${GRAPHRAG_API_KEY} # set this in the generated .env file, or remove if managed identity
model: qwen-plus
api_base: https://dashscope.aliyuncs.com/compatible-mode/v1
# api_version: 2024-05-01-preview
model_supports_json: true # recommended if this is available for your model.
concurrent_requests: 25
async_mode: threaded # or asyncio
retry_strategy: exponential_backoff
max_retries: 2
tokens_per_minute: null
requests_per_minute: null
default_embedding_model:
type: embedding
model_provider: openai
auth_type: api_key
api_key: ${GRAPHRAG_API_KEY}
model: text-embedding-v4
api_base: https://dashscope.aliyuncs.com/compatible-mode/v1
# api_version: 2024-05-01-preview
concurrent_requests: 10
async_mode: threaded # or asyncio
retry_strategy: exponential_backoff
max_retries: 2
tokens_per_minute: null
requests_per_minute: null

Input settings

input:
storage:
type: file # or blob
base_dir: "input"
file_type: text # [csv, text, json]

chunks:
size: 1200
overlap: 100
group_by_columns: [id]

Output/storage settings

If blob storage is specified in the following four sections,

connection_string and container_name must be provided

output:
type: file # [file, blob, cosmosdb]
base_dir: "output"

cache:
type: file # [file, blob, cosmosdb]
base_dir: "cache"

reporting:
type: file # [file, blob]
base_dir: "logs"

vector_store:
default_vector_store:
type: lancedb
db_uri: output/lancedb
container_name: default

Workflow settings

embed_text:
model_id: default_embedding_model
vector_store_id: default_vector_store

extract_graph:
model_id: default_chat_model
prompt: "prompts/extract_graph.txt"
entity_types: [organization,person,geo,event]
max_gleanings: 1

summarize_descriptions:
model_id: default_chat_model
prompt: "prompts/summarize_descriptions.txt"
max_length: 500

extract_graph_nlp:
text_analyzer:
extractor_type: regex_english # [regex_english, syntactic_parser, cfg]
async_mode: threaded # or asyncio

cluster_graph:
max_cluster_size: 10

extract_claims:
enabled: false
model_id: default_chat_model
prompt: "prompts/extract_claims.txt"
description: "Any claims or facts that could be relevant to information discovery."
max_gleanings: 1

community_reports:
model_id: default_chat_model
graph_prompt: "prompts/community_report_graph.txt"
text_prompt: "prompts/community_report_text.txt"
max_length: 2000
max_input_length: 8000

embed_graph:
enabled: false # if true, will generate node2vec embeddings for nodes

umap:
enabled: false # if true, will generate UMAP embeddings for nodes (embed_graph must also be enabled)

snapshots:
graphml: false
embeddings: false

Query settings

The prompt locations are required here, but each search method has a number of optional knobs that can be tuned.

See the config docs: https://microsoft.github.io/graphrag/config/yaml/#query

local_search:
chat_model_id: default_chat_model
embedding_model_id: default_embedding_model
prompt: "prompts/local_search_system_prompt.txt"

global_search:
chat_model_id: default_chat_model
map_prompt: "prompts/global_search_map_system_prompt.txt"
reduce_prompt: "prompts/global_search_reduce_system_prompt.txt"
knowledge_prompt: "prompts/global_search_knowledge_system_prompt.txt"

drift_search:
chat_model_id: default_chat_model
embedding_model_id: default_embedding_model
prompt: "prompts/drift_search_system_prompt.txt"
reduce_prompt: "prompts/drift_search_reduce_prompt.txt"

basic_search:
chat_model_id: default_chat_model
embedding_model_id: default_embedding_model
prompt: "prompts/basic_search_system_prompt.txt"

Steps to reproduce

No response

Expected Behavior

No response

GraphRAG Config Used

# Paste your config here

Logs and screenshots

No response

Additional Information

• GraphRAG Version:
• Operating System:
• Python Version:
• Related Issues:

[yelangweiwei]

call_args:
encoding_format: "float" 参数问题，可以解决以上bug

[libingtao981]

call_args: encoding_format: "float" 参数问题，可以解决以上bug

How exactly should this be configured? Could you share the relevant section from the settings.yaml file for reference?

[zzzengzhe]

same question too，i have try like this
...
default_embedding_model:
type: embedding
model_provider: openai
auth_type: api_key
api_key: ${GRAPHRAG_API_KEY}
model: text-embedding-v4

---------------------------------

call_args:
  encoding_format: "float"

---------------------------------

...

but got the same error

[libingtao981]

litellm.exceptions.BadRequestError: litellm.BadRequestError: OpenAIException - Error code: 400 - {'error': {'message': '[{'type': 'literal_error', 'loc': ('body', 'function-wrap[log_extra_fields()]', 'encoding_format'), 'msg': "Input should be 'float', 'base64' or 'bytes'", 'input': None, 'ctx': {'expected': "'float', 'base64' or 'bytes'"}}, {'type': 'missing', 'loc': ('body', 'function-wrap[log_extra_fields()]', 'messages'), 'msg': 'Field required', 'input': {'input': ['This is an LLM Embedding Test String'], 'model': 'qwen3-embedding-0.6b', 'encoding_format': None}}, {'type': 'literal_error', 'loc': ('body', 'function-wrap[log_extra_fields()]', 'encoding_format'), 'msg': "Input should be 'float', 'base64' or 'bytes'", 'input': None, 'ctx': {'expected': "'float', 'base64' or 'bytes'"}}]', 'type': 'Bad Request', 'param': None, 'code': 400}}

[libingtao981]

It has been verified that explicitly passing "encoding_format": "float" into the base_args dictionary within the _create_base_embeddings function resolves the issue, as shown below:

base_args: dict[str, Any] = {
"drop_params": True, # LiteLLM drop unsupported params for selected model.
"model": f"{model_provider}/{model}",
"timeout": model_config.request_timeout,
"api_base": model_config.api_base,
"api_version": model_config.api_version,
"api_key": model_config.api_key,
"organization": model_config.organization,
"proxy": model_config.proxy,
"audience": model_config.audience,
"encoding_format": "float",
}

[libingtao981]

Using **model_config.call_args also throws an error.

[Qwer-11-max]

embedding_models:
default_embedding_model:
model_provider: openai
model: text-embedding-v4
api_base: https://dashscope.aliyuncs.com/compatible-mode/v1
auth_method: api_key
api_key: ${GRAPHRAG_API_KEY}
call_args:
encoding_format: "float"
retry:
type: exponential_backoff

这样能解决问题

[zzzengzhe]

报了其他的一些错误

ValidationError: 2 validation errors for GraphRagConfig
models.embedding_models.type
Field required [type=missing, input_value={'default_embedding_model...ests_per_minute': None}}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.12/v/missing
models.embedding_models.model
Field required [type=missing, input_value={'default_embedding_model...ests_per_minute': None}}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.12/v/missing

[Qwer-11-max]

不知道，我将call_args加入后就可以使用了，或许是注意缩进吧，这个是embeding_model下的一个参数，格式如下
embeding_models:
--default_enbedding_model:
----call_args:
------encoding_format: "float"
除了api_base外其他参数都是默认生成的

[Qwer-11-max]

报了其他的一些错误

ValidationError: 2 validation errors for GraphRagConfig models.embedding_models.type Field required [type=missing, input_value={'default_embedding_model...ests_per_minute': None}}, input_type=dict] For further information visit https://errors.pydantic.dev/2.12/v/missing models.embedding_models.model Field required [type=missing, input_value={'default_embedding_model...ests_per_minute': None}}, input_type=dict] For further information visit https://errors.pydantic.dev/2.12/v/missing

字面意思，你的embeding_model缺少model和type参数

[zzzengzhe]

初始化的写法不是这样吗？为啥你的embedding部分开头是
embeding_models:
--default_enbedding_model:

[Qwer-11-max]

版本不一样吧，我这个是今天才安装的

…

---Original--- From: ***@***.***> Date: Mon, Feb 9, 2026 15:37 PM To: ***@***.***>; Cc: ***@***.******@***.***>; Subject: Re: [microsoft/graphrag] [Bug]: when i use qwen embedding model toindex, i get a problem (Issue #2194) zzzengzhe left a comment (microsoft/graphrag#2194) image.png (view on web) 初始化的写法不是这样吗？为啥你的embedding部分开头是 embeding_models: --default_enbedding_model: — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>