Unsloth: Cannot determine model type for config file: None
i followed the step by step guide to load the sinllama model as shown in model card. i got into below error.
TypeError Traceback (most recent call last)
/tmp/ipykernel_451/782459752.py in <cell line: 0>()
----> 1 model, _ = FastLanguageModel.from_pretrained(
2 model_name = model_name,
3 max_seq_length = max_seq_length,
4 dtype = dtype,
5 load_in_4bit = load_in_4bit,
2 frames/usr/local/lib/python3.13/dist-packages/unsloth/models/loader_utils.py in _wrapper(*args, **kwargs)
1597 _pb_were_disabled = _progress_bars_were_disabled() # restore before any retry
1598 try:
-> 1599 return fn(*args, **kwargs)
1600 except Exception as e:
1601 # Skip if not network-related, or already retried by a nested decorator
/usr/local/lib/python3.13/dist-packages/unsloth/models/loader.py in from_pretrained(model_name, max_seq_length, dtype, load_in_4bit, load_in_8bit, load_in_16bit, full_finetuning, token, device_map, device_map_planner_kwargs, rope_scaling, fix_tokenizer, trust_remote_code, use_gradient_checkpointing, resize_model_vocab, revision, use_exact_model_name, offload_embedding, float32_mixed_precision, fast_inference, gpu_memory_utilization, float8_kv_cache, random_state, max_lora_rank, disable_log_stats, qat_scheme, load_in_fp8, unsloth_tiled_mlp, text_only, *args, **kwargs)
805 raise RuntimeError(combined_error) from _cause
806
--> 807 model_types = get_transformers_model_type(
808 peft_config if peft_config is not None else model_config,
809 trust_remote_code = trust_remote_code,
/usr/local/lib/python3.13/dist-packages/unsloth_zoo/hf_utils.py in get_transformers_model_type(config, trust_remote_code)
208 # or joins the list. Treat empty and None the same.
209 if not model_types:
--> 210 raise TypeError(f"Unsloth: Cannot determine model type for config file: {str(config)}")
211 # Standardize model_type
212 final_model_types = []
TypeError: Unsloth: Cannot determine model type for config file: None
I solved my issue. the issue was my hugging face account did not have access to meta-llama/Meta-Llama-3-8B model repository. i have to request the access the y filling out their form at the front page of model repository.
I solved my issue. the issue was my hugging face account did not have access to meta-llama/Meta-Llama-3-8B model repository. i have to request the access the y filling out their form at the front page of model repository.
Yes, you have to request access to meta-llama/Meta-Llama-3-8B to load the SinLlama model.
__I am reopening this discussion because others who got the same issue can fix it by reading this discussion.