Promote RC2->main: 1M context + fused-MoE ignore regex for cross-engine load; card updates (SGLang NVFP4 note, sampling defers to generation_config, new-checkpoint note)

README.md

@@ -27,6 +27,8 @@ base_model:
 # Laguna S 2.1-NVFP4
 Laguna S 2.1-NVFP4 is a 117.6B total parameter Mixture-of-Experts model with 8.5B activated parameters per token designed for agentic coding and long-horizon work on a local machine. It uses Sliding Window Attention with per-head gating in 36 out of 48 layers for fast inference and low KV cache requirements.
 
+> **Updated release (August 2026).** This is a new checkpoint that supersedes the earlier version of this repository. The weights have changed, not only the config, so if you downloaded a previous copy please re-download to pick up the current checkpoint.
+
 
 ## Highlights
 - **Mixed SWA and global attention layout**: Laguna S 2.1 uses softplus gating with per-layer rotary scales, enabling mixed SWA (Sliding Window Attention) and global attention layers in a 3:1 ratio (across 48 total layers)
@@ -46,7 +48,7 @@ Laguna S 2.1-NVFP4 is a 117.6B total parameter Mixture-of-Experts model with 8.5
 - Experts: 256 experts with 1 shared expert
 - Sliding Window: 512 tokens
 - Modality: text-to-text
-- Context window: 262,144 tokens
+- Context window: 1,048,576 tokens
 - Reasoning support: interleaved thinking with preserved thinking
 
 
@@ -74,30 +76,30 @@ Benchmarks as of 21 July 2026. Laguna S 2.1 in **bold**; a dash (-) marks a benc
 
 ### Context length
 
-This checkpoint ships configured for a 262,144-token (256K) context window. This is the configuration we recommend for best output quality.
+This checkpoint ships configured for a 1,048,576-token (1M) context window. The weights are native 1M checkpoints: training included a long-context extension stage up to 1,048,576 tokens, and quantization was calibrated at the 1M configuration.
 
-The weights are native 1M checkpoints: training included a long-context extension stage up to 1,048,576 tokens, and quantization was calibrated at the 1M configuration. If you need more than 256K of context, restore the 1M configuration by editing `config.json`:
+If you prefer to cap the context at 262,144 tokens (256K), edit `config.json`:
 
 ```json
 "rope_parameters": {
   "full_attention": {
-    "factor": 128.0,
-    "attention_factor": 1.4852030263919618
+    "factor": 32.0,
+    "attention_factor": 1.3465735902799727
   }
 },
-"max_position_embeddings": 1048576
+"max_position_embeddings": 262144
 ```
 
-You may experience quality degradation with the 1M configuration. If you use it, we recommend sampling with `temperature <= 0.7` and `top_p <= 0.95`.
+At long context you may experience some quality degradation.
 
-### Recommended sampling
+### Sampling
 
-For the best balance of quality and reliability we recommend sampling with `temperature 0.7` and `top_p 0.95`.
+The sampling defaults in the checkpoint's `generation_config.json` are authoritative (`top_k 20` is eval-certified truncation). Serve with those defaults rather than setting a separate temperature or top_p.
 
 
 ### Local deployment
 
-Laguna S 2.1-NVFP4 is supported in vLLM, SGLang and Transformers, and TRT-LLM thanks to the support of the team at NVIDIA. Use Laguna-S 2.1 with Ollama (with MLX support) or Llama.cpp (BF16 and Q4\_K\_M only) for the best results on your local machine.
+Laguna S 2.1-NVFP4 is supported in vLLM, Transformers, and TRT-LLM thanks to the support of the team at NVIDIA. NVFP4 does not currently run correctly on SGLang (see the SGLang note below). Use Laguna-S 2.1 with Ollama (with MLX support) or Llama.cpp (BF16 and Q4\_K\_M only) for the best results on your local machine.
 
 #### vLLM
 
@@ -179,7 +181,6 @@ vllm serve poolside/Laguna-S-2.1-NVFP4 \
   --enable-auto-tool-choice \
   --tool-call-parser poolside_v1 \
   --reasoning-parser poolside_v1 \
-  --override-generation-config '{"temperature":0.7,"top_p":0.95}' \
   --max-num-seqs 32 \
   --max-model-len 262144 \
   --gpu-memory-utilization 0.85 \
@@ -189,12 +190,11 @@ vllm serve poolside/Laguna-S-2.1-NVFP4 \
 - You do not need backend flags: auto-selection picks FlashInferCutlass, which
   runs natively on `sm_121`. Do not set `--linear-backend flashinfer_b12x` on
   0.25.1; the opt-in is broken there and it is slower anyway.
-- Keep the `--override-generation-config`. Many clients send no sampling
-  parameters, and the raw defaults degrade output on the NVFP4 quantization. The
-  model's `generation_config.json` sets `top_k 20` (eval-certified truncation),
-  which handles that. Do not add `min_p`: vLLM rejects `min_p` and `logit_bias`
-  under speculative decoding, so putting it in the defaults returns a 400 on
-  every sampled request.
+- Sampling comes from the checkpoint's `generation_config.json` (`top_k 20`
+  eval-certified truncation); serve with those defaults rather than overriding
+  them. Do not add `min_p`: vLLM rejects `min_p` and `logit_bias` under
+  speculative decoding, so putting it in the defaults returns a 400 on every
+  sampled request.
 - `--max-num-seqs 32` is required: DFlash crashes vLLM at the default of 256.
 - The first start takes about 15 minutes (weight load from NVMe, JIT, and graph
   capture).
@@ -229,7 +229,9 @@ network.
 
 #### SGLang
 
-Laguna S 2.1 is supported in SGLang via [sgl-project/sglang#24204](https://github.com/sgl-project/sglang/pull/24204). Quantization is detected automatically from `quantization_config`, so no extra flags are required. See the [SGLang cookbook entry](https://docs.sglang.io/cookbook/autoregressive/Poolside/Laguna-S-2.1) and the main [Laguna S 2.1 model card](https://huggingface.co/poolside/Laguna-S-2.1) for a serving recipe.
+> **NVFP4 is not currently working on SGLang.** On Blackwell GPUs the model loads but its NVFP4 (W4A4) fused-MoE kernel produces NaN and degenerate output, so generations are corrupted. This is a SGLang engine issue, not a checkpoint problem: the same weights serve correctly on vLLM and TRT-LLM, and we have a fix in progress upstream. For SGLang today, serve the [FP8 checkpoint](https://huggingface.co/poolside/Laguna-S-2.1-FP8) instead.
+
+The Laguna S 2.1 architecture is supported in SGLang via [sgl-project/sglang#24204](https://github.com/sgl-project/sglang/pull/24204). Quantization is detected automatically from `quantization_config`, so no extra flags are required. See the [SGLang cookbook entry](https://docs.sglang.io/cookbook/autoregressive/Poolside/Laguna-S-2.1) and the main [Laguna S 2.1 model card](https://huggingface.co/poolside/Laguna-S-2.1) for a serving recipe.
 
 #### Transformers
 

config.json

@@ -2,61 +2,72 @@
   "architectures": [
     "LagunaForCausalLM"
   ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
   "auto_map": {
     "AutoConfig": "configuration_laguna.LagunaConfig",
     "AutoModelForCausalLM": "modeling_laguna.LagunaForCausalLM"
   },
-  "model_type": "laguna",
-  "vocab_size": 100352,
-  "hidden_size": 3072,
-  "intermediate_size": 12288,
-  "num_hidden_layers": 48,
-  "num_attention_heads": 48,
-  "num_key_value_heads": 8,
-  "head_dim": 128,
-  "max_position_embeddings": 262144,
-  "attention_bias": false,
-  "attention_dropout": 0.0,
-  "rms_norm_eps": 1e-06,
-  "num_experts": 256,
-  "num_experts_per_tok": 10,
-  "moe_intermediate_size": 1024,
-  "shared_expert_intermediate_size": 1024,
-  "norm_topk_prob": true,
-  "router_aux_loss_coef": 0.0,
-  "moe_router_logit_softcapping": 0.0,
-  "decoder_sparse_step": 1,
-  "mlp_only_layers": [
-    0
-  ],
   "bos_token_id": 2,
+  "decoder_sparse_step": 1,
   "eos_token_id": [
     2,
     24
   ],
-  "pad_token_id": 9,
-  "tie_word_embeddings": false,
-  "use_cache": true,
-  "torch_dtype": "bfloat16",
   "gating": "per-head",
-  "sliding_window": 512,
-  "rope_parameters": {
-    "full_attention": {
-      "rope_theta": 500000.0,
-      "rope_type": "yarn",
-      "factor": 32.0,
-      "original_max_position_embeddings": 8192,
-      "beta_slow": 1.0,
-      "beta_fast": 32.0,
-      "attention_factor": 1.3465735902799727,
-      "partial_rotary_factor": 0.5
-    },
-    "sliding_attention": {
-      "rope_type": "default",
-      "rope_theta": 10000.0,
-      "partial_rotary_factor": 1.0
-    }
-  },
+  "gating_types": [
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head",
+    "per_head"
+  ],
+  "head_dim": 128,
+  "hidden_size": 3072,
+  "intermediate_size": 12288,
   "layer_types": [
     "full_attention",
     "sliding_attention",
@@ -107,7 +118,7 @@
     "sliding_attention",
     "sliding_attention"
   ],
-  "moe_apply_router_weight_on_input": false,
+  "max_position_embeddings": 1048576,
   "mlp_layer_types": [
     "dense",
     "sparse",
@@ -158,57 +169,16 @@
     "sparse",
     "sparse"
   ],
-  "gating_types": [
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head",
-    "per_head"
+  "mlp_only_layers": [
+    0
   ],
+  "model_type": "laguna",
+  "moe_apply_router_weight_on_input": false,
+  "moe_intermediate_size": 1024,
   "moe_routed_scaling_factor": 2.5,
+  "moe_router_logit_softcapping": 0.0,
+  "norm_topk_prob": true,
+  "num_attention_heads": 48,
   "num_attention_heads_per_layer": [
     48,
     72,
@@ -259,6 +229,11 @@
     72,
     72
   ],
+  "num_experts": 256,
+  "num_experts_per_tok": 10,
+  "num_hidden_layers": 48,
+  "num_key_value_heads": 8,
+  "pad_token_id": 9,
   "quantization_config": {
     "config_groups": {
       "group_0": {
@@ -269,7 +244,7 @@
           "dynamic": "local",
           "group_size": 16,
           "num_bits": 4,
-          "observer": "minmax",
+          "observer": "static_minmax",
           "observer_kwargs": {},
           "scale_dtype": "torch.float8_e4m3fn",
           "strategy": "tensor_group",
@@ -279,7 +254,6 @@
         },
         "output_activations": null,
         "targets": [
-          "re:.*experts\\.[0-9]+\\.(gate_proj|up_proj|down_proj)$",
           "re:.*experts\\.[0-9]+\\.(gate_proj|up_proj|down_proj)$"
         ],
         "weights": {
@@ -302,37 +276,49 @@
     "global_compression_ratio": null,
     "ignore": [
       "lm_head",
-      "model.layers.0.mlp.gate_proj",
-      "model.layers.0.mlp.up_proj",
-      "model.layers.0.mlp.down_proj",
       "re:.*\\.self_attn\\.q_proj$",
       "re:.*\\.self_attn\\.k_proj$",
       "re:.*\\.self_attn\\.v_proj$",
       "re:.*\\.self_attn\\.o_proj$",
       "re:.*\\.self_attn\\.g_proj$",
       "re:.*\\.mlp\\.gate$",
+      "model.layers.0.mlp.gate_proj",
+      "model.layers.0.mlp.up_proj",
+      "model.layers.0.mlp.down_proj",
       "re:.*\\.mlp\\.shared_expert\\.gate_proj$",
       "re:.*\\.mlp\\.shared_expert\\.up_proj$",
-      "re:.*\\.mlp\\.shared_expert\\.down_proj$"
+      "re:.*\\.mlp\\.shared_expert\\.down_proj$",
+      "re:^model\\.layers\\.4[0-7]\\.mlp\\.experts(\\..*)?$"
     ],
-    "kv_cache_scheme": {
-      "actorder": null,
-      "block_structure": null,
-      "dynamic": false,
-      "group_size": null,
-      "num_bits": 8,
-      "observer": "minmax",
-      "observer_kwargs": {},
-      "scale_dtype": null,
-      "strategy": "tensor",
-      "symmetric": true,
-      "type": "float",
-      "zp_dtype": null
-    },
     "quant_method": "compressed-tensors",
     "quantization_status": "compressed",
     "sparsity_config": {},
     "transform_config": {},
     "version": "0.14.1.dev11+gf2ee47b"
-  }
+  },
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "full_attention": {
+      "attention_factor": 1.4852030263919618,
+      "beta_fast": 32.0,
+      "beta_slow": 1.0,
+      "factor": 128.0,
+      "original_max_position_embeddings": 8192,
+      "partial_rotary_factor": 0.5,
+      "rope_theta": 500000.0,
+      "rope_type": "yarn"
+    },
+    "sliding_attention": {
+      "partial_rotary_factor": 1.0,
+      "rope_theta": 10000.0,
+      "rope_type": "default"
+    }
+  },
+  "router_aux_loss_coef": 0.0,
+  "shared_expert_intermediate_size": 1024,
+  "sliding_window": 512,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "use_cache": true,
+  "vocab_size": 100352
 }

configuration_laguna.py

@@ -12,16 +12,9 @@
 # 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.
-from transformers.configuration_utils import PretrainedConfig as PreTrainedConfig
-try:
-    from transformers.modeling_rope_utils import RopeParameters
-except ImportError:
-    RopeParameters = dict
-try:
-    from transformers.utils.import_utils import is_causal_conv1d_available, is_flash_linear_attention_available
-except ImportError:
-    def is_causal_conv1d_available(): return False
-    def is_flash_linear_attention_available(): return False
+from transformers.configuration_utils import PreTrainedConfig
+from transformers.modeling_rope_utils import RopeParameters
+from transformers.utils.import_utils import is_causal_conv1d_available, is_flash_linear_attention_available
 
 
 class LagunaConfig(PreTrainedConfig):

generation_config.json

@@ -8,17 +8,17 @@
   "pad_token_id": 9,
   "temperature": 1.0,
   "top_p": 1.0,
-  "speculative_config": {
-    "method": "dflash",
-    "source": "huggingface",
-    "model": "poolside/Laguna-S-2.1-DFlash-NVFP4",
-    "num_speculative_tokens": 15
-  },
+  "min_p": 0.0,
+  "top_k": 20,
   "tool_call_parser": "poolside_v1",
   "reasoning_parser": "poolside_v1",
   "default_chat_template_kwargs": {
     "enable_thinking": true
   },
-  "top_k": 20,
-  "min_p": 0.0
+  "speculative_config": {
+    "method": "dflash",
+    "source": "huggingface",
+    "model": "poolside/Laguna-S-2.1-DFlash",
+    "num_speculative_tokens": 15
+  }
 }

model-00015-of-00015.safetensors → model-00001-of-00049.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd81d64092e4da18e97492afdfade1d1cd1b8e46e2fcc1c623e8c183dee6da2b
-size 233635672
+oid sha256:bafa5ef7ccfde8cd63db88f74f9461aba9035370b3c86ccb754fc6f2ba010041
+size 509720000

model-00002-of-00015.safetensors → model-00002-of-00049.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:29769ce86e021d92d788ed7c9f35e5ce62ea9a0ac46df50352fd0128cb258afc
-size 5120328828
+oid sha256:d91f55c3c81b5aadec0b826ee817e8b410d0b659c045898f1c84ba9e5d167725
+size 588356928

model-00001-of-00015.safetensors → model-00003-of-00049.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8591d5d36f8dcc6acf60dfdb57b1ba11ee133a3a19de1d4f3ffad0ae6e43c60
-size 5116269068
+oid sha256:80e0783d4e23806821eb5e1f5240230fd5c509e071a883270949f051ec3ada99
+size 2982209384

model-00003-of-00015.safetensors → model-00004-of-00049.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c4143771eece564da2e8f2166dcd5cde38f692691c624473794904d3449669f
-size 5120530032
+oid sha256:e3ee5dbe50a75da5776ff2e5b94b846e80964a083582696ffbbd3b4912e5050d
+size 5115055792

model-00005-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5a6c393d8ab17abdbd4c28f2656c56d998f6d2d123ea2e97e8816b8ec918e19f
-size 5120542216

model-00004-of-00015.safetensors → model-00005-of-00049.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4460d1b13f24e9ef0aac71aed1479c90f55b997180592a0b86433ad2aef96e1
-size 5120531224
+oid sha256:88bb23dc4e9a56b102db5a7e96a0151521b7680eb3df783c83a32bfca203284e
+size 5115055792

model-00006-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:209cbcb8960e9f2987fc780beac5f2f7db7216c8b265e38a9c0fb153d386b273
-size 5120541608

model-00006-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44d5cc53e65d78f86ef8fca5f51af06eae98bf24a9084545bd007bd24d6509f9
+size 5115055792

model-00007-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c0597b0d9ce14215f1393c4be305402901328a9e517b850a9ed2deba14310574
-size 5120541600

model-00007-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1a371ec20caa5311788f59b4650f540f5792128e499f6d84317d99eab765b4b
+size 5115055792

model-00008-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f2b57043b744e36cc07d06d4e9051226736d1ececabe259a424868040515795a
-size 5120542112

model-00008-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df485c3c09d7f0f2d3748b9db400dd53b0cd8b952e7e560e563b2992bb415354
+size 5115055792

model-00009-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b10777318e89e836ee63ee408e71c9fd9268e62fc7e2e312b8ab3288f51d86bb
-size 5120542424

model-00009-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3460f9cd677c6ba9e7f7d06d474a19dd1bcfd8605e4688a9cb6010f8329e1810
+size 5115055792

model-00010-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c609513bc62fb1c0e1b06765252e806fcef9fd28a533c53b656c47ff3c9c0311
-size 5120541600

model-00010-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5143801097f26d20e8f1f30d8b9fc1f38bc3ccae9e5fb37e12e31a7f4c9a0f71
+size 4982932568

model-00011-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bde2636d5f14a7883768002dec05b81d8efeffad8499a55ad5dce8b61b611ce6
-size 5120541608

model-00011-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50ed03d6c8ead6938c4dd21e39aa6ee61892ebfca2db74181284bc81d3ac702c
+size 4570085320

model-00012-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c4ddbd51880dbbf0d93e479866124ce3038205c9012467c6869630d13b0d33e2
-size 5120541904

model-00012-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3b0c9afde987635b22a7777046188c1946beefb01115bb2e7ff69392bc8d884
+size 2970115792

model-00013-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6a5d7c9e0023da16e297b577d2ef2611019ee23aef687b5c2cab7bd3fe98684d
-size 5120542312

model-00013-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83f91b41d1bc72b0fe4c9d0dc7f6ce241a1e2da8692483afec16246c1082eaef
+size 1438983752

model-00014-of-00015.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7e1fb566bfb4315c9f38977f02efddc1cb15b465fcdd24c590a59e28f1644684
-size 5120541912

model-00014-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7335bd2c696ed2ac8bb1b7aa007c805c10ad35dfb8adcc02a6dae48ef6cb8e9d
+size 1438983752

model-00015-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:419dbbcd9abad0c5f70bda0c592dba5e852a4d526b0573585b76e871ed3af074
+size 1438983752

model-00016-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bc0b91dd99412d36811950e8727b3e48aad0dece6ff025aae3aed59b6b0c6b7
+size 1438983752

model-00017-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33295ddcae4328168e2aac3f36457afa8058e966f7738a78c6d9239727ad808c
+size 1438983752

model-00018-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86262f28e4be8b57fde0931d3a5b1c138309127cba1afc3dce23e701f8ce9969
+size 1438983752

model-00019-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12d70338ab96a189cc5eecc328e07562bfd07a80ea660eff708c87d06750d1d6
+size 1438983752

model-00020-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6f56b5670234436dc949940ddfb86a471690d5081620a47e462d44ab6086f23
+size 1438983752

model-00021-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec304cbc979e3606446ba236a3ae577324c6e51f91d94f3855b989bb0b0e437a
+size 1438986752

model-00022-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2dc96ac8104cea542bfbd0eacbbfbaf041b7f7b1ef3c852fcaab0bed1f467ee6
+size 1438986784

model-00023-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:039a3eeedd0a5f5985eca6a052202e3872259b21adbe0319ec2cf3f0df39fae1
+size 1438986808

model-00024-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caf75a44fa15c0c79900b171e0051a64a3a40b1b5213d0019bd6e2269bb541fe
+size 1438986824

model-00025-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c4e33932c116d3ffc2f9187536c778617c1da53b356068fc32617ff57beee7e
+size 1438986824

model-00026-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1399e19f57ce3c1b315055457564ed89673d7ae1a34a2847209fda20047ec26c
+size 1438986824

model-00027-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4e9fdc200593f301dc0ebe5c973ad4d3d72940ecec9521eb6e2e09f2f1ac000
+size 1438986824

model-00028-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce15771b0aebb9d930cdc64aabec9f7857dc6c5697daa072242e74e129e53d32
+size 1438986824

model-00029-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16a2c631593187c0ee53fa8dd5e5ed13eb906102cefc42ca2c8b8cf9eb87654d
+size 1438986992

model-00030-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b580ce6dd5346afa4c797054c6d1953e3a8d64bfe56824d67d116894c19953ca
+size 1438987000

model-00031-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38f197c2cdb310b796e9159ae94c3cff7aafc9b4d308292561f93042eed61ded
+size 1438987000

model-00032-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2e6dbb74fb87cdad3efa291735c2b9e8cc087846aa957b07dd6e00530f45373
+size 1438987000

model-00033-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05b86a9e41d488ebfc276a2cfa95196bd9938205fdd0768c4f093a91447eaf34
+size 1438987000

model-00034-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4b17618337501b7bfa1f8491e3d07db5f92e4dfa0eefe2597f5df705b6b0226
+size 1438987000

model-00035-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:355e7e0a26497c4a23a63ce25e8d191f84abd6ee589fdd140599c96bc8b25e20
+size 1438987000

model-00036-of-00049.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:634998dddce9d1418b802f9718f92cab512dee939fef1c5c35fe8a17bc05e427
+size 1438987000