[ UX ] List of Modifiers Not Working Properly

[robertgshaw2-neuralmagic]

• Running with the following is not working properly
• Running with the recipe is working properly

from datasets import load_dataset
from transformers import AutoTokenizer

from llmcompressor.modifiers.quantization import GPTQModifier
from llmcompressor.modifiers.smoothquant import SmoothQuantModifier
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot

# Select model and load it.
MODEL_ID = "meta-llama/Meta-Llama-3-8B-Instruct"
model = SparseAutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    device_map="auto",
    torch_dtype="auto",
)
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)

# Select calibration dataset.
DATASET_ID = "HuggingFaceH4/ultrachat_200k"
DATASET_SPLIT = "train_sft"

# Select number of samples. 512 samples is a good place to start.
# Increasing the number of samples can improve accuracy.
NUM_CALIBRATION_SAMPLES = 512
MAX_SEQUENCE_LENGTH = 2048

# Load dataset and preprocess.
ds = load_dataset(DATASET_ID, split=DATASET_SPLIT)
ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES))

def preprocess(example):
    return {
        "text": tokenizer.apply_chat_template(
            example["messages"],
            tokenize=False,
        )
    }

ds = ds.map(preprocess)

# Tokenize inputs.
def tokenize(sample):
    return tokenizer(
        sample["text"],
        padding=False,
        max_length=MAX_SEQUENCE_LENGTH,
        truncation=True,
        add_special_tokens=False,
    )

ds = ds.map(tokenize, remove_columns=ds.column_names)

# Configure algorithms. In this case, we:
#   * apply SmoothQuant to make the activations easier to quantize
#   * quantize the weights to int8 with GPTQ (static per channel)
#   * quantize the activations to int8 (dynamic per token)
# Note: set sequential_update: true in the recipe to reduce memory
recipe = [
    SmoothQuantModifier(smoothing_strength=0.8),
    GPTQModifier(targets="Linear", scheme="W8A8", ignore=["lm_head"]),
]

# Apply algorithms.
oneshot(
    model=model,
    dataset=ds,
    recipe=recipe,
    max_seq_length=MAX_SEQUENCE_LENGTH,
    num_calibration_samples=NUM_CALIBRATION_SAMPLES,
)

# Confirm generations of the quantized model look sane.
print("\n\n")
print("========== SAMPLE GENERATION ==============")
input_ids = tokenizer("Hello my name is", return_tensors="pt").input_ids.to("cuda")
output = model.generate(input_ids, max_new_tokens=100)
print(tokenizer.decode(output[0]))
print("==========================================\n\n")

# Save to disk compressed.
SAVE_DIR = MODEL_ID.split("/")[1] + "-W8A8-Dynamic-Per-Token"
model.save_pretrained(SAVE_DIR, save_compressed=True)
tokenizer.save_pretrained(SAVE_DIR)

[robertgshaw2-neuralmagic]

quant_stage:
    quant_modifiers:
        SmoothQuantModifier:
            smoothing_strength: 0.8
            mappings: [
                [["re:.*q_proj", "re:.*k_proj", "re:.*v_proj"], "re:.*input_layernorm"],
                [["re:.*gate_proj", "re:.*up_proj"], "re:.*post_attention_layernorm"]
            ]
        GPTQModifier:
            # Set sequential_update: true to reduce memory consumption
            sequential_update: false
            ignore: ["lm_head"]
            config_groups:
                group_0:
                    weights:
                        num_bits: 8
                        type: "int"
                        symmetric: true
                        strategy: "channel"
                    input_activations:
                        num_bits: 8
                        type: "int"
                        symmetric: true
                        dynamic: true
                        strategy: "token"
                    targets: ["Linear"]

[robertgshaw2-neuralmagic]

completed