Fixing gemma2.

server/text_generation_server/models/__init__.py

@@ -68,6 +68,9 @@ try:
     from text_generation_server.models.flash_gemma import (
         FlashGemma,
     )
+    from text_generation_server.models.flash_gemma2 import (
+        FlashGemma2,
+    )
     from text_generation_server.models.pali_gemma import (
         PaliGemma,
     )
@@ -102,6 +105,7 @@ if FLASH_ATTENTION:
     __all__.append(FlashQwen2)
     __all__.append(FlashStarcoder2)
     __all__.append(FlashGemma)
+    __all__.append(FlashGemma2)
     __all__.append(FlashCohere)
 
 MAMBA_AVAILABLE = True
@@ -143,6 +147,11 @@ class ModelType(enum.Enum):
         "name": "Gemma",
         "url": "https://huggingface.co/google/gemma-7b",
     }
+    GEMMA2 = {
+        "type": "gemma2",
+        "name": "Gemma2",
+        "url": "https://huggingface.co/google/gemma2-9b",
+    }
     COHERE = {
         "type": "cohere",
         "name": "Cohere",
@@ -630,6 +639,27 @@ def get_model(
                 dtype=dtype,
                 trust_remote_code=trust_remote_code,
             )
+    elif model_type == GEMMA2:
+        if FLASH_ATTENTION:
+            return FlashGemma2(
+                model_id,
+                revision,
+                quantize=quantize,
+                speculator=speculator,
+                dtype=dtype,
+                trust_remote_code=trust_remote_code,
+            )
+        elif sharded:
+            raise NotImplementedError(FLASH_ATT_ERROR_MESSAGE.format("Sharded Gemma2"))
+        else:
+            return CausalLM(
+                model_id,
+                revision,
+                quantize=quantize,
+                speculator=speculator,
+                dtype=dtype,
+                trust_remote_code=trust_remote_code,
+            )
 
     if model_type == COHERE:
         if FLASH_ATTENTION:

server/text_generation_server/models/custom_modeling/flash_gemma2_modeling.py

@@ -0,0 +1,500 @@
+# coding=utf-8
+# Copyright 2022 EleutherAI and the HuggingFace Inc. team. All rights reserved.
+#
+# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
+# and OPT implementations in this library. It has been modified from its
+# original forms to accommodate minor architectural differences compared
+# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import torch
+import torch.distributed
+
+from torch import nn
+from transformers.activations import ACT2FN
+from transformers.configuration_utils import PretrainedConfig
+from typing import Optional, List, Tuple
+
+from text_generation_server.layers.attention import (
+    paged_attention,
+    attention,
+    reshape_and_cache,
+)
+from text_generation_server.layers import (
+    TensorParallelRowLinear,
+    TensorParallelColumnLinear,
+    TensorParallelEmbedding,
+    SpeculativeHead,
+    get_linear,
+)
+from text_generation_server.layers.rotary import PositionRotaryEmbedding
+from text_generation_server.layers.layernorm import (
+    FastRMSNorm,
+)
+
+
+class Gemma2Config(PretrainedConfig):
+    def __init__(
+        self,
+        vocab_size=256128,
+        hidden_size=3072,
+        intermediate_size=24576,
+        num_hidden_layers=28,
+        num_attention_heads=16,
+        num_key_value_heads=16,
+        head_dim=256,
+        hidden_act="gelu_pytorch_tanh",
+        max_position_embeddings=8192,
+        initializer_range=0.02,
+        rms_norm_eps=1e-6,
+        use_cache=True,
+        pad_token_id=None,
+        bos_token_id=1,
+        eos_token_id=2,
+        tie_word_embeddings=True,
+        rope_theta=10000.0,
+        rope_scaling=None,
+        attention_bias=False,
+        attention_dropout=0.0,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.max_position_embeddings = max_position_embeddings
+        self.hidden_size = hidden_size
+        self.head_dim = head_dim
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+
+        # for backward compatibility
+        if num_key_value_heads is None:
+            num_key_value_heads = num_attention_heads
+
+        self.num_key_value_heads = num_key_value_heads
+        self.hidden_act = hidden_act
+        self.initializer_range = initializer_range
+        self.rms_norm_eps = rms_norm_eps
+        self.use_cache = use_cache
+        self.rope_theta = rope_theta
+        self.rope_scaling = rope_scaling
+        self.attention_bias = attention_bias
+        self.attention_dropout = attention_dropout
+
+        super().__init__(
+            pad_token_id=pad_token_id,
+            bos_token_id=bos_token_id,
+            eos_token_id=eos_token_id,
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )
+
+
+class Gemma2FastRMSNorm(FastRMSNorm):
+    @classmethod
+    def load(cls, prefix, weights, eps=1e-6):
+        dtype = weights.dtype
+        weights.dtype = torch.float32
+        weight = weights.get_tensor(f"{prefix}.weight") + 1
+        weights.dtype = dtype
+        new = cls(weight, eps)
+        new.dtype = dtype
+        return new
+
+    # perform the multiplication in full precision and downcast after
+    def forward(self, hidden_states, residual=None):
+        if residual is not None:
+            hidden_states += residual
+        residual = hidden_states
+        hidden_states = hidden_states.to(torch.float32)
+        variance = hidden_states.pow(2).mean(-1, keepdim=True)
+        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
+        hidden_states = hidden_states * self.weight
+        return hidden_states.to(self.dtype), residual
+
+
+def load_attention(config, prefix, weights):
+    if config.num_attention_heads != config.num_key_value_heads:
+        return _load_gqa(config, prefix, weights)
+    else:
+        return TensorParallelColumnLinear.load_multi(
+            config,
+            prefixes=[f"{prefix}.q_proj", f"{prefix}.k_proj", f"{prefix}.v_proj"],
+            dim=0,
+            weights=weights,
+            bias=False,
+        )
+
+
+def _load_gqa(config, prefix: str, weights):
+    assert config.num_attention_heads % weights.process_group.size() == 0
+
+    weight = weights.get_multi_weights_col(
+        prefixes=[f"{prefix}.q_proj", f"{prefix}.k_proj", f"{prefix}.v_proj"],
+        quantize=config.quantize,
+        dim=0,
+    )
+
+    if config.quantize not in ["gptq", "awq", "marlin"]:
+        weight = weight.to(dtype=weights.dtype).to(device=weights.device)
+
+        head_size = config.head_dim
+        num_heads = config.num_attention_heads // weights.process_group.size()
+        num_key_value_heads = config.num_key_value_heads // weights.process_group.size()
+        assert list(weight.shape) == [
+            (num_heads + 2 * num_key_value_heads) * head_size,
+            config.hidden_size,
+        ], f"{list(weight.shape)} != {[(num_heads + 2 * config.num_key_value_heads) * head_size, config.hidden_size]}"
+
+    return TensorParallelColumnLinear(
+        get_linear(weight, bias=None, quantize=config.quantize)
+    )
+
+
+class FlashGemma2Attention(torch.nn.Module):
+    def __init__(self, prefix: str, config, weights, causal: bool, is_sliding: bool):
+        super().__init__()
+        self.num_heads = config.num_attention_heads
+        self.head_size = config.head_dim
+        self.causal = causal
+        if is_sliding:
+            self.window_size = config.sliding_window
+        else:
+            self.window_size = -1
+
+        self.rotary_emb = PositionRotaryEmbedding.static(
+            config=config,
+            dim=self.head_size,
+            base=config.rope_theta,
+            device=weights.device,
+        )
+
+        # self.softmax_scale = self.head_size**-0.5
+        self.softmax_scale = config.query_pre_attn_scalar**-0.5
+
+        if self.num_heads % weights.process_group.size() != 0:
+            raise ValueError(
+                f"`num_heads` must be divisible by `num_shards` (got `num_heads`: {self.num_heads} "
+                f"and `num_shards`: {weights.process_group.size()}"
+            )
+        self.num_heads = self.num_heads // weights.process_group.size()
+        self.num_key_value_heads = (
+            config.num_key_value_heads // weights.process_group.size()
+        )
+
+        self.query_key_value = load_attention(config, prefix, weights)
+
+        self.o_proj = TensorParallelRowLinear.load(
+            config,
+            prefix=f"{prefix}.o_proj",
+            weights=weights,
+            bias=False,
+        )
+        self.num_groups = self.num_heads // self.num_key_value_heads
+        self.kv_head_mapping = torch.arange(
+            0, self.num_key_value_heads, dtype=torch.int32, device=weights.device
+        ).repeat_interleave(self.num_groups)
+
+    def forward(
+        self,
+        hidden_states,
+        cos,
+        sin,
+        cu_seqlen_prefill,
+        kv_cache,
+        block_tables,
+        slots,
+        input_lengths,
+        max_s,
+    ):
+        qkv = self.query_key_value(hidden_states)
+        query, kv = qkv.split(
+            [
+                self.head_size * self.num_heads,
+                2 * self.head_size * self.num_key_value_heads,
+            ],
+            dim=1,
+        )
+        query = query.view(-1, self.num_heads, self.head_size)
+        kv = kv.view(-1, 2, self.num_key_value_heads, self.head_size)
+
+        self.rotary_emb(query, torch.select(kv, dim=1, index=0), cos, sin)
+
+        reshape_and_cache(kv[:, 0], kv[:, 1], kv_cache[0], kv_cache[1], slots)
+
+        # output tensor
+        attn_output = torch.empty_like(query)
+
+        # Prefill
+        if cu_seqlen_prefill is not None:
+            # flash attention
+            attention(
+                query,
+                torch.select(kv, dim=1, index=0),
+                torch.select(kv, dim=1, index=1),
+                attn_output,
+                cu_seqlen_prefill,
+                max_s,
+                self.softmax_scale,
+                causal=self.causal,
+                window_size_left=self.window_size,
+            )
+        # Decode
+        else:
+            paged_attention(
+                attn_output,
+                query,
+                kv_cache[0],
+                kv_cache[1],
+                self.kv_head_mapping,
+                self.softmax_scale,
+                block_tables,
+                input_lengths,
+                max_s,
+            )
+
+        return self.o_proj(attn_output.view(-1, self.num_heads * self.head_size))
+
+
+class Gemma2MLP(nn.Module):
+    def __init__(self, prefix, config, weights):
+        super().__init__()
+        act = config.hidden_act
+        self.act = (
+            ACT2FN[act]
+            if "gelu" not in act
+            else lambda x: torch.nn.functional.gelu(
+                x,
+                approximate=(
+                    "tanh" if act in ["gelu_fast", "gelu_pytorch_tanh"] else "none"
+                ),
+            )
+        )
+        # Fuse gate and up proj
+        self.gate_up_proj = TensorParallelColumnLinear.load_multi(
+            config,
+            prefixes=[f"{prefix}.gate_proj", f"{prefix}.up_proj"],
+            weights=weights,
+            dim=0,
+            bias=False,
+        )
+        self.down_proj = TensorParallelRowLinear.load(
+            config,
+            prefix=f"{prefix}.down_proj",
+            weights=weights,
+            bias=False,
+        )
+        self.intermediate_size = (
+            config.intermediate_size // weights.process_group.size()
+        )
+
+    def forward(self, hidden_states):
+        gate_up_states = self.gate_up_proj(hidden_states)
+        gate_up_states = gate_up_states.view(-1, 2, self.intermediate_size)
+        return self.down_proj(self.act(gate_up_states[:, 0]) * gate_up_states[:, 1])
+
+
+class FlashGemma2Layer(nn.Module):
+    def __init__(self, prefix, config, weights, causal: bool, is_sliding: bool):
+        super().__init__()
+        self.self_attn = FlashGemma2Attention(
+            prefix=f"{prefix}.self_attn",
+            config=config,
+            weights=weights,
+            causal=causal,
+            is_sliding=is_sliding,
+        )
+        self.mlp = Gemma2MLP(prefix=f"{prefix}.mlp", config=config, weights=weights)
+
+        self.input_layernorm = Gemma2FastRMSNorm.load(
+            prefix=f"{prefix}.input_layernorm", weights=weights, eps=config.rms_norm_eps
+        )
+        self.post_attention_layernorm = Gemma2FastRMSNorm.load(
+            prefix=f"{prefix}.post_attention_layernorm",
+            weights=weights,
+            eps=config.rms_norm_eps,
+        )
+        self.pre_feedforward_layernorm = Gemma2FastRMSNorm.load(
+            prefix=f"{prefix}.pre_feedforward_layernorm",
+            weights=weights,
+            eps=config.rms_norm_eps,
+        )
+        self.post_feedforward_layernorm = Gemma2FastRMSNorm.load(
+            prefix=f"{prefix}.post_feedforward_layernorm",
+            weights=weights,
+            eps=config.rms_norm_eps,
+        )
+
+    def forward(
+        self,
+        hidden_states,
+        residual,
+        cos,
+        sin,
+        cu_seqlen_prefill,
+        kv_cache,
+        block_tables,
+        slots,
+        input_lengths,
+        max_s,
+    ):
+        normed_hidden_states, res = self.input_layernorm(hidden_states, residual)
+
+        # Self Attention
+        attn_output = self.self_attn(
+            normed_hidden_states,
+            cos,
+            sin,
+            cu_seqlen_prefill,
+            kv_cache,
+            block_tables,
+            slots,
+            input_lengths,
+            max_s,
+        )
+
+        # faster post attention rms norm
+        normed_attn_res_output, _ = self.post_attention_layernorm(attn_output)
+        normed_attn_res_output = normed_attn_res_output + res
+        res = normed_attn_res_output
+
+        pre_normed, _ = self.pre_feedforward_layernorm(normed_attn_res_output)
+        mlp_output = self.mlp(pre_normed)
+        post_hidden_states, _ = self.post_feedforward_layernorm(mlp_output)
+
+        return post_hidden_states, normed_attn_res_output
+
+
+class FlashGemma2Model(torch.nn.Module):
+    def __init__(self, prefix, config, weights, causal: bool):
+        super().__init__()
+
+        process_group = weights.process_group
+        self.tp_rank = process_group.rank()
+        self.tp_world_size = process_group.size()
+        self.layers = nn.ModuleList(
+            [
+                FlashGemma2Layer(
+                    prefix=f"{prefix}.layers.{layer_id}",
+                    config=config,
+                    weights=weights,
+                    causal=causal,
+                    is_sliding=layer_id % 2 == 0,
+                )
+                for layer_id in range(config.num_hidden_layers)
+            ]
+        )
+        self.norm = Gemma2FastRMSNorm.load(
+            prefix=f"{prefix}.norm", weights=weights, eps=config.rms_norm_eps
+        )
+
+        self.head_size = self.layers[0].self_attn.head_size
+        self.num_heads = self.layers[0].self_attn.num_heads
+        self.num_key_value_heads = self.layers[0].self_attn.num_key_value_heads
+
+    def forward(
+        self,
+        inputs_embeds: torch.Tensor,
+        position_ids: torch.Tensor,
+        cu_seqlen_prefill: Optional[torch.Tensor],
+        kv_cache: List[Tuple[torch.Tensor, torch.Tensor]],
+        block_tables: torch.Tensor,
+        slots: torch.Tensor,
+        input_lengths: torch.Tensor,
+        max_s: int,
+    ) -> torch.Tensor:
+        hidden_states = inputs_embeds
+
+        # Get rotary cos and sin for this forward
+        # Avoid to index in each layer
+        cos, sin = self.layers[0].self_attn.rotary_emb.get_cos_sin(
+            position_ids, max_s, hidden_states.dtype
+        )
+
+        residual = None
+        for i, layer in enumerate(self.layers):
+            hidden_states, residual = layer(
+                hidden_states,
+                residual,
+                cos,
+                sin,
+                cu_seqlen_prefill,
+                kv_cache[i],
+                block_tables,
+                slots,
+                input_lengths,
+                max_s,
+            )
+
+        hidden_states, _ = self.norm(hidden_states, residual)
+
+        return hidden_states
+
+
+class FlashGemma2ForCausalLM(torch.nn.Module):
+    def __init__(self, prefix, config, weights, causal: bool):
+        super().__init__()
+
+        embed_norm = config.hidden_size**0.5
+        if not prefix:
+            prefix = "model"
+        else:
+            prefix = f"{prefix}.model"
+
+        self.embed_tokens = TensorParallelEmbedding(
+            prefix=f"{prefix}.embed_tokens", weights=weights
+        )
+        self.embed_tokens.weight *= embed_norm
+
+        self.model = FlashGemma2Model(
+            prefix=prefix, config=config, weights=weights, causal=causal
+        )
+        self.lm_head = SpeculativeHead.load(
+            prefix=(
+                f"{prefix}.embed_tokens"
+                if config.tie_word_embeddings
+                else f"{prefix}.lm_head"
+            ),
+            config=config,
+            weights=weights,
+        )
+
+    def forward(
+        self,
+        input_ids: torch.Tensor,
+        position_ids: torch.Tensor,
+        cu_seqlen_prefill: Optional[torch.Tensor],
+        kv_cache: List[Tuple[torch.Tensor, torch.Tensor]],
+        block_tables: torch.Tensor,
+        slots: torch.Tensor,
+        input_lengths: torch.Tensor,
+        max_s: int,
+        prefill_cache_indices: Optional[torch.Tensor],
+        lm_head_indices: Optional[torch.Tensor] = None,
+        adapter_data: Optional[torch.Tensor] = None,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
+        input_embeds = self.embed_tokens(input_ids)
+        hidden_states = self.model(
+            input_embeds,
+            position_ids,
+            cu_seqlen_prefill,
+            kv_cache,
+            block_tables,
+            slots,
+            input_lengths,
+            max_s,
+        )
+        if lm_head_indices is not None:
+            hidden_states = hidden_states[lm_head_indices]
+        logits, speculative_logits = self.lm_head(hidden_states)
+        return logits, speculative_logits

server/text_generation_server/models/custom_modeling/flash_gemma_modeling.py

@@ -375,8 +375,6 @@ class FlashGemmaModel(torch.nn.Module):
             prefix=f"{prefix}.norm", weights=weights, eps=config.rms_norm_eps
         )
 
-        self.gradient_checkpointing = False
-
         self.head_size = self.layers[0].self_attn.head_size
         self.num_heads = self.layers[0].self_attn.num_heads
         self.num_key_value_heads = self.layers[0].self_attn.num_key_value_heads

server/text_generation_server/models/flash_causal_lm.py

@@ -28,8 +28,12 @@ from text_generation_server.models.types import (
     GeneratedText,
 )
 from text_generation_server.pb import generate_pb2
-from text_generation_server.models.globals import MEM_POOL, CUDA_GRAPHS
-import text_generation_server.models.globals as tgi_globals
+from text_generation_server.models.globals import (
+    MEM_POOL,
+    CUDA_GRAPHS,
+    get_adapter_to_index,
+    MODEL_ID,
+)
 from text_generation_server.utils import StoppingCriteria, HeterogeneousNextTokenChooser
 from text_generation_server.utils.dist import MEMORY_FRACTION
 from text_generation_server.utils.segments import SegmentConcatBuilder, find_segments
@@ -233,7 +237,8 @@ class FlashCausalLMBatch(Batch):
             stopping_criterias.append(stopping_criteria)
             top_n_tokens.append(r.top_n_tokens)
 
-            adapter_index = tgi_globals.ADAPTER_TO_INDEX.get(r.adapter_id, 0)
+            ADAPTER_TO_INDEX = get_adapter_to_index()
+            adapter_index = ADAPTER_TO_INDEX.get(r.adapter_id, 0)
             adapter_indices_list.append(torch.full((input_length,), adapter_index))
             adapter_set.add(adapter_index)
 
@@ -499,9 +504,8 @@ class FlashCausalLMBatch(Batch):
 
             top_n_tokens.append(self.top_n_tokens[idx])
 
-            adapter_index = tgi_globals.ADAPTER_TO_INDEX.get(
-                self.requests[idx].adapter_id, 0
-            )
+            ADAPTER_TO_INDEX = get_adapter_to_index()
+            adapter_index = ADAPTER_TO_INDEX.get(self.requests[idx].adapter_id, 0)
             adapter_set.add(adapter_index)
 
             remaining_tokens = (
@@ -1017,7 +1021,7 @@ class FlashCausalLM(Model):
 
                 tunableop_filepath = os.path.join(
                     HUGGINGFACE_HUB_CACHE,
-                    f"tunableop_{tgi_globals.MODEL_ID.replace('/', '-')}_tp{self.world_size}_rank{self.rank}.csv",
+                    f"tunableop_{MODEL_ID.replace('/', '-')}_tp{self.world_size}_rank{self.rank}.csv",
                 )
 
                 logger.info(

server/text_generation_server/models/flash_gemma2.py

@@ -0,0 +1,75 @@
+import torch
+import torch.distributed
+
+from opentelemetry import trace
+from typing import Optional
+from transformers import PretrainedConfig, AutoTokenizer
+
+from text_generation_server.models import FlashCausalLM
+from text_generation_server.models.custom_modeling.flash_gemma2_modeling import (
+    FlashGemma2ForCausalLM,
+)
+from text_generation_server.utils import (
+    initialize_torch_distributed,
+    weight_files,
+    Weights,
+)
+
+tracer = trace.get_tracer(__name__)
+
+
+class FlashGemma2(FlashCausalLM):
+    def __init__(
+        self,
+        model_id: str,
+        revision: Optional[str] = None,
+        quantize: Optional[str] = None,
+        speculator: Optional[str] = None,
+        dtype: Optional[torch.dtype] = None,
+        trust_remote_code: bool = False,
+    ):
+        self.process_group, rank, world_size = initialize_torch_distributed()
+        if torch.cuda.is_available():
+            device = torch.device(f"cuda:{rank}")
+            dtype = torch.bfloat16 if dtype is None else dtype
+        else:
+            raise NotImplementedError("FlashGemma2 is only available on GPU")
+
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_id,
+            revision=revision,
+            padding_side="left",
+            truncation_side="left",
+            trust_remote_code=trust_remote_code,
+        )
+
+        config = PretrainedConfig.from_pretrained(
+            model_id, revision=revision, trust_remote_code=trust_remote_code
+        )
+        config.quantize = quantize
+        config.speculator = speculator
+
+        torch.distributed.barrier(group=self.process_group)
+
+        filenames = weight_files(model_id, revision=revision, extension=".safetensors")
+        weights = Weights(filenames, device, dtype, process_group=self.process_group)
+        if config.quantize in ["gptq", "awq", "marlin"]:
+            weights._set_gptq_params(model_id, revision)
+
+        # TODO hardcoded
+        prefix = ""
+        model = FlashGemma2ForCausalLM(prefix, config, weights, causal=True)
+
+        torch.distributed.barrier(group=self.process_group)
+        super(FlashGemma2, self).__init__(
+            model_id=model_id,
+            model=model,
+            tokenizer=tokenizer,
+            num_layers=len(model.model.layers),
+            num_kv_heads=model.model.num_key_value_heads,
+            head_size=model.model.head_size,
+            dtype=dtype,
+            device=device,
+            rank=rank,
+            world_size=world_size,
+        )

server/text_generation_server/models/globals.py

@@ -44,3 +44,8 @@ ADAPTER_TO_INDEX: Dict[str, int] = None
 def set_adapter_to_index(adapter_to_index: Dict[str, int]):
     global ADAPTER_TO_INDEX
     ADAPTER_TO_INDEX = adapter_to_index
+
+
+def get_adapter_to_index():
+    global ADAPTER_TO_INDEX
+    return ADAPTER_TO_INDEX