"llama_model_load: error loading model: check_tensor_dims: tensor 'token_embd.weight' not found" after computing AWQ scales and applying them to the gguf model

[Autism-al]

When I quantified the Qwen2.5-1.5B-instruct model according to "GGUF Export" in the examples.md in the docs, it showed that the quantization was complete and I obtained the gguf model.But when I load the model through llama-cpp-python, the following error was reported:

llama_model_load: error loading model: check_tensor_dims: tensor 'token_embd.weight' not found

After multiple attempts, I have confirmed that the model obtained by directly quantifying with llama.cpp without using autoawq is normal, and I can correctly read it by using llama.cpp.the original model Qwen2.5-1.5B-instruct can also use transformers for inference normally.

This is my quantization code:

import os
import subprocess
from awq import AutoAWQForCausalLM
from transformers import AutoTokenizer

model_path = 'Qwen2.5-1.5B-Instruct'
quant_path = 'Qwen2.5-1.5B-Instruct-awq'
llama_cpp_path = 'llama.cpp'
quant_config = { "zero_point": True, "q_group_size": 128, "w_bit": 4, "version": "GEMM" }

# Load model
model = AutoAWQForCausalLM.from_pretrained(
    model_path, low_cpu_mem_usage=True, use_cache=False, device_map="cuda",
)
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)

# Quantize
# NOTE: We avoid packing weights, so you cannot use this model in AutoAWQ
# after quantizing. The saved model is FP16 but has the AWQ scales applied.
model.quantize(
    tokenizer,
    quant_config=quant_config,
    export_compatible=True
)

# Save quantized model
model.save_quantized(quant_path)
tokenizer.save_pretrained(quant_path)
print(f'Model is quantized and saved at "{quant_path}"')

# GGUF conversion
print('Converting model to GGUF...')
llama_cpp_method = "q4_K_M"
convert_cmd_path = os.path.join(llama_cpp_path, "convert_hf_to_gguf.py")
quantize_cmd_path = os.path.join(llama_cpp_path, "llama-quantize")

if not os.path.exists(llama_cpp_path):
    cmd = f"git clone https://github.com/ggerganov/llama.cpp.git {llama_cpp_path} && cd {llama_cpp_path} && make LLAMA_CUBLAS=1 LLAMA_CUDA_F16=1"
    subprocess.run([cmd], shell=True, check=True)

subprocess.run([
    f"python {convert_cmd_path} {quant_path} --outfile {quant_path}/slu-int4.gguf"
], shell=True, check=True)

subprocess.run([
    f"{quantize_cmd_path} {quant_path}/slu-int4.gguf {quant_path}/slu_{llama_cpp_method}.gguf {llama_cpp_method}"
], shell=True, check=True)

It's interesting that before this, I had been using this way(awq+convert to Q4K gguf) for quantization, and I could always read it correctly by using llama.cpp. Recently, when I changed the conda environment, this problem began to occur. Is it due to the impact of version changes? I have been troubled by this problem for several days now, but I really can't restore to the original conda environment.

The complete error log is as follows：
llama_model_loader: loaded meta data with 26 key-value pairs and 338 tensors from Qwen2.5-1.5B-Instruct-awq/slu-int4.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = qwen2
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Qwen2.5 1.5B Instruct
llama_model_loader: - kv 3: general.finetune str = Instruct
llama_model_loader: - kv 4: general.basename str = Qwen2.5
llama_model_loader: - kv 5: general.size_label str = 1.5B
llama_model_loader: - kv 6: qwen2.block_count u32 = 28
llama_model_loader: - kv 7: qwen2.context_length u32 = 32768
llama_model_loader: - kv 8: qwen2.embedding_length u32 = 1536
llama_model_loader: - kv 9: qwen2.feed_forward_length u32 = 8960
llama_model_loader: - kv 10: qwen2.attention.head_count u32 = 12
llama_model_loader: - kv 11: qwen2.attention.head_count_kv u32 = 2
llama_model_loader: - kv 12: qwen2.rope.freq_base f32 = 1000000.000000
llama_model_loader: - kv 13: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001
llama_model_loader: - kv 14: general.file_type u32 = 1
llama_model_loader: - kv 15: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 16: tokenizer.ggml.pre str = qwen2
llama_model_loader: - kv 17: tokenizer.ggml.tokens arr[str,151936] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 19: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151645
llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643
llama_model_loader: - kv 22: tokenizer.ggml.bos_token_id u32 = 151643
llama_model_loader: - kv 23: tokenizer.ggml.add_bos_token bool = false
llama_model_loader: - kv 24: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>...
llama_model_loader: - kv 25: general.quantization_version u32 = 2
llama_model_loader: - type f32: 141 tensors
llama_model_loader: - type f16: 197 tensors
llm_load_vocab: special tokens cache size = 22
llm_load_vocab: token to piece cache size = 0.9310 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = qwen2
llm_load_print_meta: vocab type = BPE
llm_load_print_meta: n_vocab = 151936
llm_load_print_meta: n_merges = 151387
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 32768
llm_load_print_meta: n_embd = 1536
llm_load_print_meta: n_layer = 28
llm_load_print_meta: n_head = 12
llm_load_print_meta: n_head_kv = 2
llm_load_print_meta: n_rot = 128
llm_load_print_meta: n_swa = 0
llm_load_print_meta: n_embd_head_k = 128
llm_load_print_meta: n_embd_head_v = 128
llm_load_print_meta: n_gqa = 6
llm_load_print_meta: n_embd_k_gqa = 256
llm_load_print_meta: n_embd_v_gqa = 256
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-06
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 8960
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 2
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 32768
llm_load_print_meta: rope_finetuned = unknown
llm_load_print_meta: ssm_d_conv = 0
llm_load_print_meta: ssm_d_inner = 0
llm_load_print_meta: ssm_d_state = 0
llm_load_print_meta: ssm_dt_rank = 0
llm_load_print_meta: ssm_dt_b_c_rms = 0
llm_load_print_meta: model type = ?B
llm_load_print_meta: model ftype = F16
llm_load_print_meta: model params = 1.54 B
llm_load_print_meta: model size = 2.88 GiB (16.00 BPW)
llm_load_print_meta: general.name = Qwen2.5 1.5B Instruct
llm_load_print_meta: BOS token = 151643 '<|endoftext|>'
llm_load_print_meta: EOS token = 151645 '<|im_end|>'
llm_load_print_meta: PAD token = 151643 '<|endoftext|>'
llm_load_print_meta: LF token = 148848 'ÄĬ'
llm_load_print_meta: EOT token = 151645 '<|im_end|>'
llm_load_print_meta: max token length = 256
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: yes
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 4070, compute capability 8.9, VMM: yes
llm_load_tensors: ggml ctx size = 0.15 MiB
llama_model_load: error loading model: check_tensor_dims: tensor 'token_embd.weight' not found
llama_load_model_from_file: failed to load model
Traceback (most recent call last):
File "/home/lmf/llm/Qwen2.5-finetuning/llama-cpp-instruct.py", line 2, in
llm = Llama(model_path="Qwen2.5-1.5B-Instruct-awq/slu-int4.gguf")
File "/home/lmf/anaconda3/envs/qwenslu/lib/python3.10/site-packages/llama_cpp/llama.py", line 371, in init
_LlamaModel(
File "/home/lmf/anaconda3/envs/qwenslu/lib/python3.10/site-packages/llama_cpp/_internals.py", line 55, in init
raise ValueError(f"Failed to load model from file: {path_model}")
ValueError: Failed to load model from file: Qwen2.5-1.5B-Instruct-awq/slu-int4.gguf
ERROR conda.cli.main_run:execute(124): conda run python /home/lmf/llm/Qwen2.5-finetuning/llama-cpp-instruct.py failed. (See above for error)

[Ivy233]

I encountered the same problem in Qwen2-1.5B and Qwen2.5-3B. Both found the weight "embed_tokens", but did not find its existence when exporting to gguf.